fix(egress): prevent stale active-endpoint overwrite

Make the endpoint watcher the sole active-endpoint annotation writer so a Service snapshot captured during startup cannot replace a newer endpoint. Preserve intentionally empty snapshots across endpoint-to-zero-to-same transitions, select the cleared annotation from the configured egress family, reject malformed or wrong-family endpoints, and preserve the other family in dual-stack state.

Fixes #1673.

Signed-off-by: Marcel Fest <marcel.fest@telekom.de>

.github/workflows/anchore-syft.yml

@@ -22,10 +22,10 @@ jobs:
     runs-on: ubuntu-latest
     steps:
       - name: Checkout
-        uses: actions/checkout@v4
+        uses: actions/checkout@v7
         with:
           ref: ${{ github.ref_name }}
       - name: Anchore SBOM Action
-        uses: anchore/sbom-action@v0.20.4
+        uses: anchore/sbom-action@v0.24.0
         with:
           format: cyclonedx-json

.github/workflows/ci-pull-request.yaml

@@ -9,43 +9,63 @@ jobs:
       matrix:
         mode: ["arp", "rt", "bgp"]
       fail-fast: true
-      max-parallel: 1
     steps:
+      - name: Get current date
+        id: date
+        run: echo "::set-output name=date::$(date +'%Y-%m-%d-%H-%M')"
       - name: Ensure fs wont cause issues
         run: sudo sysctl fs.inotify.max_user_instances=8192 &&  sudo sysctl fs.inotify.max_user_watches=524288
       - name: Checkout code
-        uses: actions/checkout@v4
+        uses: actions/checkout@v7
       - name: Install Go
-        uses: actions/setup-go@v5
+        uses: actions/setup-go@v7
         with:
           go-version-file: go.mod
       - name: Build image locally
         run: make dockerx86Local
       - name: Run Manifest generation tests
         run: make manifest-test
-      - name: Run ARP mode tests v1.29.0 onwards
-        run: make e2e-tests129-arp
+      - name: Run ARP mode tests
+        run: DOCKER_API_VERSION=1.48 E2E_KEEP_LOGS=true make e2e-tests-arp
         if: matrix.mode== 'arp'
-      - name: Run RT mode tests v1.29.0 onwards
-        run: make e2e-tests129-rt
+      - name: Run RT mode tests
+        run: DOCKER_API_VERSION=1.48 E2E_KEEP_LOGS=true make e2e-tests-rt
         if: matrix.mode== 'rt'
       - name: Get GoBGP binaries
         run: make get-gobgp
         if: matrix.mode== 'bgp'
-      - name: Run BGP mode tests v1.29.0 onwards
-        run: sudo -E PATH=$PATH DOCKER_API_VERSION=1.48 make e2e-tests129-bgp
+      - name: Run BGP mode tests
+        run: sudo -E PATH=$PATH DOCKER_API_VERSION=1.48 E2E_KEEP_LOGS=true make e2e-tests-bgp
         if: matrix.mode== 'bgp'
+      - name: Change log directory permissions
+        run: sudo chmod -R 755 /tmp/kube-vip-test*
+        if: matrix.mode== 'bgp' && always()
+      - name: Save logs
+        uses: actions/upload-artifact@v7
+        with:
+          name: e2e-test-logs-${{ matrix.mode }}-${{ steps.date.outputs.date }}
+          path: /tmp/kube-vip-test*
+        if: always()
   service-e2e-tests:
     runs-on: ubuntu-latest
     name: E2E service tests
     steps:
+      - name: Get current date
+        id: date
+        run: echo "::set-output name=date::$(date +'%Y-%m-%d-%H-%M')"
       - name: Checkout code
-        uses: actions/checkout@v4
+        uses: actions/checkout@v7
       - name: Install Go
-        uses: actions/setup-go@v5
+        uses: actions/setup-go@v7
         with:
           go-version-file: go.mod
       - name: Build image with iptables
         run: make dockerx86ActionIPTables
       - name: Run tests
-        run: DOCKERTAG=action make service-tests
+        run: DOCKER_API_VERSION=1.48 DOCKERTAG=action E2E_KEEP_LOGS=true make service-tests
+      - name: Save logs
+        uses: actions/upload-artifact@v7
+        with:
+          name: services-test-logs-${{ steps.date.outputs.date }}
+          path: /tmp/kube-vip-services*
+        if: always()

.github/workflows/ci.yaml

@@ -10,11 +10,14 @@ jobs:
       - name: Init
         run: sudo apt-get update && sudo apt-get install -y build-essential &&  sudo sysctl fs.inotify.max_user_instances=8192 &&  sudo sysctl fs.inotify.max_user_watches=524288
       - name: Install golangci-lint
-        run: curl -sSfL https://raw.githubusercontent.com/golangci/golangci-lint/master/install.sh | sh -s -- -b $(go env GOPATH)/bin v1.64.8
+        uses: golangci/golangci-lint-action@v9
+        with:
+          version: v2.12
+          install-only: true
       - name: Checkout code
-        uses: actions/checkout@v4
+        uses: actions/checkout@v7
       - name: Install Go
-        uses: actions/setup-go@v5
+        uses: actions/setup-go@v7
         with:
           go-version-file: go.mod
       - name: All checks
@@ -24,9 +27,9 @@ jobs:
     name: Unit tests
     steps:
       - name: Checkout code
-        uses: actions/checkout@v4
+        uses: actions/checkout@v7
       - name: Install Go
-        uses: actions/setup-go@v5
+        uses: actions/setup-go@v7
         with:
           go-version-file: go.mod
       - name: Run tests
@@ -36,9 +39,9 @@ jobs:
     runs-on: ubuntu-latest
     steps:
       - name: Checkout code
-        uses: actions/checkout@v4
+        uses: actions/checkout@v7
       - name: Install Go
-        uses: actions/setup-go@v5
+        uses: actions/setup-go@v7
         with:
           go-version-file: go.mod
       - name: Run tests
@@ -48,9 +51,9 @@ jobs:
     name: Image vulnerability scan
     steps:
       - name: Checkout code
-        uses: actions/checkout@v4
+        uses: actions/checkout@v7
       - name: Build image with iptables
-        run: make dockerx86ActionIPTables
+        run: make dockerx86Action
       - name: Run Trivy vulnerability scanner
         uses: aquasecurity/trivy-action@master
         with:

.github/workflows/codeql-analysis.yml

@@ -38,16 +38,16 @@ jobs:
 
     steps:
     - name: Checkout repository
-      uses: actions/checkout@v4
+      uses: actions/checkout@v7
       
     - name: Install Go
-      uses: actions/setup-go@v5
+      uses: actions/setup-go@v7
       with:
         go-version-file: go.mod
         
     # Initializes the CodeQL tools for scanning.
     - name: Initialize CodeQL
-      uses: github/codeql-action/init@v3
+      uses: github/codeql-action/init@v4
       with:
         languages: ${{ matrix.language }}
         # If you wish to specify custom queries, you can do so here or in a config file.
@@ -58,7 +58,7 @@ jobs:
     # Autobuild attempts to build any compiled languages  (C/C++, C#, or Java).
     # If this step fails, then you should remove it and run the build manually (see below)
     - name: Autobuild
-      uses: github/codeql-action/autobuild@v3
+      uses: github/codeql-action/autobuild@v4
 
     # ℹ️ Command-line programs to run using the OS shell.
     # 📚 https://git.io/JvXDl
@@ -72,4 +72,4 @@ jobs:
     #   make release
 
     - name: Perform CodeQL Analysis
-      uses: github/codeql-action/analyze@v3
+      uses: github/codeql-action/analyze@v4

.github/workflows/main.yaml

@@ -11,25 +11,25 @@ jobs:
     runs-on: ubuntu-latest
     steps:
       - name: Checkout
-        uses: actions/checkout@v4
+        uses: actions/checkout@v7
       - name: Set up QEMU
-        uses: docker/setup-qemu-action@v3
+        uses: docker/setup-qemu-action@v4
       - name: Set up Docker Buildx
-        uses: docker/setup-buildx-action@v3
+        uses: docker/setup-buildx-action@v4
       - name: Login to DockerHub
-        uses: docker/login-action@v3
+        uses: docker/login-action@v4
         with:
           username: ${{ secrets.DOCKERHUB_USERNAME }}
           password: ${{ secrets.DOCKERHUB_TOKEN }}
       - name: Login to Github Packages
-        uses: docker/login-action@v3
+        uses: docker/login-action@v4
         with:
           registry: ghcr.io
           username: ${{ github.actor }}
           password: ${{ secrets.GITHUB_TOKEN }}   
       - name: Build standard version
         id: docker_build
-        uses: docker/build-push-action@v6
+        uses: docker/build-push-action@v7
         with:
           context: .
           platforms: linux/amd64,linux/arm/v7,linux/arm64,linux/ppc64le,linux/s390x
@@ -39,7 +39,7 @@ jobs:
             ghcr.io/kube-vip/kube-vip:${{ github.ref_name }}
       - name: Build iptables version
         id: docker_build_iptables
-        uses: docker/build-push-action@v6
+        uses: docker/build-push-action@v7
         with:
           context: .
           file: Dockerfile_iptables

.github/workflows/release.yaml

@@ -11,31 +11,31 @@ jobs:
     runs-on: ubuntu-latest
     steps:
       - name: Checkout
-        uses: actions/checkout@v4
+        uses: actions/checkout@v7
       - name: Set up QEMU
-        uses: docker/setup-qemu-action@v3
+        uses: docker/setup-qemu-action@v4
       - name: Set up Docker Buildx
-        uses: docker/setup-buildx-action@v3
+        uses: docker/setup-buildx-action@v4
       - name: Login to DockerHub
-        uses: docker/login-action@v3
+        uses: docker/login-action@v4
         with:
           username: ${{ secrets.DOCKERHUB_USERNAME }}
           password: ${{ secrets.DOCKERHUB_TOKEN }}
       - name: Login to Github Packages
-        uses: docker/login-action@v3
+        uses: docker/login-action@v4
         with:
           registry: ghcr.io
           username: ${{ github.actor }}
           password: ${{ secrets.GITHUB_TOKEN }}
       - name: Generate Metadata
-        uses: docker/metadata-action@v5.7.0
+        uses: docker/metadata-action@v6.2.0
         id: metadata
         with:
           labels: |
             org.opencontainers.image.documentation=https://kube-vip.io/docs/
       - name: Build and push main branch
         id: docker_build
-        uses: docker/build-push-action@v6
+        uses: docker/build-push-action@v7
         with:
           context: .
           platforms: linux/amd64,linux/arm/v7,linux/arm64,linux/ppc64le,linux/s390x
@@ -48,7 +48,7 @@ jobs:
             ghcr.io/kube-vip/kube-vip:latest
       - name: Build iptables version and push main branch
         id: docker_build_iptables
-        uses: docker/build-push-action@v6
+        uses: docker/build-push-action@v7
         with:
           context: .
           file: Dockerfile_iptables

.gitignore

@@ -5,3 +5,4 @@ bin
 testing/e2e/etcd/certs
 pkg/etcd/etcd.pid
 pkg/etcd/etcd-data
+testing/e2e/e2e.test

.golangci.yml

@@ -1,13 +1,38 @@
-run:
-  timeout: 10m
-  
+version: "2"
 linters:
   enable:
-  - bodyclose
-  - gofmt
-  - goimports
-  - revive
-  - gosec
-  - misspell
-  - unconvert
-  - unparam
+    - bodyclose
+    - gosec
+    - misspell
+    - unconvert
+    - unparam
+  settings:
+    misspell:
+      ignore-rules:
+        - creater
+    staticcheck:
+      checks:
+        - all
+        # Disable QF1008 to retain embedded fields for better readability.
+        - "-QF1008"
+  exclusions:
+    generated: lax
+    presets:
+      - comments
+      - common-false-positives
+      - legacy
+      - std-error-handling
+    paths:
+      - third_party$
+      - builtin$
+      - examples$
+formatters:
+  enable:
+    - gofmt
+    - goimports
+  exclusions:
+    generated: lax
+    paths:
+      - third_party$
+      - builtin$
+      - examples$

CHANGELOG.md

@@ -0,0 +1,60 @@
+# Changelog
+
+All notable changes to this project will be documented in this file.
+
+The format is based on [Keep a Changelog](https://keepachangelog.com/en/1.0.0/),
+and this project adheres to [Semantic Versioning](https://semver.org/spec/v2.0.0.html).
+
+## [Unreleased]
+
+### Fixed
+- Add a configurable kube-vip instance name and use it to isolate internal nftables egress tables, persist table ownership on Services, and migrate per-Service chains without affecting other deployments. Fixes #1634.
+- Retry on 403 Forbidden and 401 Unauthorized in `ServicesWatcher` at startup with exponential backoff. Fixes #1464.
+- Reintroduce BGP config via node annotations. Fixes #1488.
+- Fail fast in runtime `manager` and `service` paths when legacy `vip_address` is used without `vip_subnet` in control-plane ARP, BGP, or Routing Table mode.
+- Cancel the mode context on init or configuration failure before waiting on goroutines during shutdown.
+
+
+### Added
+- Configurable control-plane health check for BGP mode without leader election
+  - Polls a configurable HTTP(S) endpoint (e.g. `https://localhost:6443/livez`) to verify the exposed service is healthy (usually the local kube-apiserver)
+  - Withdraws the BGP route after a configurable number of consecutive failures, removing the unhealthy node from the ECMP set
+  - Re-announces the route automatically once the endpoint recovers
+  - Gracefully withdraws the route on shutdown (SIGTERM)
+  - Supports custom CA certificates for TLS verification
+  - Configuration via environment variables or CLI flags:
+    - `control_plane_health_check_address` / `--controlPlaneHealthCheckAddress`: URL to poll
+    - `control_plane_health_check_period_seconds` / `--controlPlaneHealthCheckPeriodSeconds`: interval between checks (default: 5)
+    - `control_plane_health_check_timeout_seconds` / `--controlPlaneHealthCheckTimeoutSeconds`: per-request timeout (default: 3)
+    - `control_plane_health_check_failure_threshold` / `--controlPlaneHealthCheckFailureThreshold`: consecutive failures before withdrawal (default: 3)
+    - `control_plane_health_check_ca_path` / `--controlPlaneHealthCheckCAPath`: CA cert for HTTPS verification
+- SIGUSR1 signal handler for runtime configuration dumps (#1301)
+  - Send SIGUSR1 to kube-vip process to dump current configuration to stdout
+  - Configuration dump includes:
+    - Basic configuration (VIP, interface, port, namespace settings)
+    - BGP configuration (enabled status, AS number, router ID, peers)
+    - ARP/NDP configuration (enabled status, broadcast rate)
+    - Services configuration (enabled status, load balancer settings)
+    - Network interfaces status
+    - Leader election configuration (type, lease details)
+    - Runtime statistics (load balancer, Prometheus, health check settings)
+  - Output format: Human-readable plaintext via fmt.Printf()
+  - Thread-safe implementation using mutex protection
+  - Non-disruptive: Process continues running after configuration dump
+  - Added comprehensive unit tests for all dump methods
+  - Added E2E tests for signal handling
+- Opt-in support for endpointless `LoadBalancer` services with `externalTrafficPolicy: Cluster`
+  - Annotation: `kube-vip.io/allow-reconcile-without-endpoints: "true"`
+  - Starts service handling path for opted-in endpointless Cluster services while preserving default endpoint-gated behavior for non-opt-in services and `Local` policy
+  - Added endpoint behavior tests and README usage documentation
+  - Added support in ipoib interfaces in ARP mode. Fixes #694
+
+### Changed
+- Updated signal handlers in manager_arp.go, manager_bgp.go, manager_wireguard.go, and manager_table.go to use switch statement pattern for handling multiple signals (SIGUSR1, SIGINT, SIGTERM)
+- wireguard.go now manages a complete wireguard interface on the current network namespace
+- manager_wireguard.go uses the new wireguard.go implementation
+
+## [v1.0.1] - Previous Release
+
+### Previous changes
+- See git history for changes prior to CHANGELOG.md introduction

CONTRIBUTING.md

@@ -113,8 +113,8 @@ and *merged* sorts of commits.
 To make it easier for reviewers to review your PR, consider the following:
 
 1. Follow the golang [coding conventions](https://github.com/golang/go/wiki/CodeReviewComments).
-2. Format your code with `make golangci-fix`; if the [linters](ci/README.md) flag an issue that
-   cannot be fixed automatically, an error message will be displayed so you can address the issue.
+2. Format your code with `make simplify` to automatically fix formatting issues.
+2. Lint your code with `make check`; if the linters flag an issue that cannot be fixed automatically, an error message will be displayed so you can address the issue.
 3. Follow [git commit](https://chris.beams.io/posts/git-commit/) guidelines.
 4. Follow [logging](https://github.com/kubernetes/community/blob/master/contributors/devel/sig-instrumentation/logging.md) guidelines.
 

Dockerfile

@@ -1,6 +1,6 @@
 # syntax=docker/dockerfile:experimental
 
-FROM golang:1.24.5-alpine3.22 as dev
+FROM golang:1.26.6-alpine3.23 as dev
 RUN apk add --no-cache git ca-certificates make
 RUN adduser -D appuser
 COPY . /src/

Dockerfile_iptables

@@ -1,6 +1,6 @@
 # syntax=docker/dockerfile:experimental
 
-FROM golang:1.24.5-alpine3.22 as dev
+FROM golang:1.26.6-alpine3.23 as dev
 RUN apk add --no-cache git make
 RUN adduser -D appuser
 COPY . /src/
@@ -11,7 +11,7 @@ RUN --mount=type=cache,sharing=locked,id=gomod,target=/go/pkg/mod/cache \
     --mount=type=cache,sharing=locked,id=goroot,target=/root/.cache/go-build \
     CGO_ENABLED=0 GOOS=linux make build
 
-FROM alpine:3.22.1
+FROM alpine:3.24.1
 # Update pkgs and add iptables
 RUN apk upgrade && \
     apk add --no-cache iptables iptables-legacy

Makefile

@@ -5,7 +5,7 @@ TARGET := kube-vip
 .DEFAULT_GOAL := $(TARGET)
 
 # These will be provided to the target
-VERSION := v1.0.0
+VERSION := v1.2.3
 
 BUILD := `git rev-parse HEAD`
 
@@ -15,9 +15,11 @@ TARGETOS=linux
 # Use linker flags to provide version/build settings to the target
 LDFLAGS=-ldflags "-s -w -X=main.Version=$(VERSION) -X=main.Build=$(BUILD) -extldflags -static"
 DOCKERTAG ?= $(VERSION)
-REPOSITORY ?= plndr
+REPOSITORY ?= docker.io/plndr
+GO_VERSION := 1.25.6
+K8S_VERSION ?= v1.35.0
 
-.PHONY: all build clean install uninstall simplify check run e2e-tests
+.PHONY: all build clean install uninstall simplify check run e2e-tests unit-tests integration-tests unit-tests-docker integration-tests-docker
 
 all: check install
 
@@ -109,45 +111,46 @@ run: install
 manifests:
 	@make build
 	@mkdir -p ./docs/manifests/$(VERSION)/
-	@./kube-vip manifest pod --interface eth0 --vip 192.168.0.1 --arp --leaderElection --controlplane --services > ./docs/manifests/$(VERSION)/kube-vip-arp.yaml
-	@./kube-vip manifest pod --interface eth0 --vip 192.168.0.1 --arp --leaderElection --controlplane --services --enableLoadBalancer > ./docs/manifests/$(VERSION)/kube-vip-arp-lb.yaml
-	@./kube-vip manifest pod --interface eth0 --vip 192.168.0.1 --bgp --controlplane --services > ./docs/manifests/$(VERSION)/kube-vip-bgp.yaml
-	@./kube-vip manifest daemonset --interface eth0 --vip 192.168.0.1 --arp --leaderElection --controlplane --services --inCluster > ./docs/manifests/$(VERSION)/kube-vip-arp-ds.yaml
-	@./kube-vip manifest daemonset --interface eth0 --vip 192.168.0.1 --arp --leaderElection --controlplane --services --inCluster --enableLoadBalancer > ./docs/manifests/$(VERSION)/kube-vip-arp-ds-lb.yaml
-	@./kube-vip manifest daemonset --interface eth0 --vip 192.168.0.1 --bgp --leaderElection --controlplane --services --inCluster > ./docs/manifests/$(VERSION)/kube-vip-bgp-ds.yaml
-	@./kube-vip manifest daemonset --interface eth0 --vip 192.168.0.1 --bgp --leaderElection --controlplane --services --inCluster > ./docs/manifests/$(VERSION)/kube-vip-bgp-em-ds.yaml
+	@./kube-vip manifest pod --interface eth0 --vip 192.168.0.1 --image "$(REPOSITORY)/$(TARGET):$(DOCKERTAG)" --arp --leaderElection --controlplane --services > ./docs/manifests/$(VERSION)/kube-vip-arp.yaml
+	@./kube-vip manifest pod --interface eth0 --vip 192.168.0.1 --image "$(REPOSITORY)/$(TARGET):$(DOCKERTAG)" --arp --leaderElection --controlplane --services --enableLoadBalancer > ./docs/manifests/$(VERSION)/kube-vip-arp-lb.yaml
+	@./kube-vip manifest pod --interface eth0 --vip 192.168.0.1 --image "$(REPOSITORY)/$(TARGET):$(DOCKERTAG)" --bgp --controlplane --services > ./docs/manifests/$(VERSION)/kube-vip-bgp.yaml
+	@./kube-vip manifest daemonset --interface eth0 --vip 192.168.0.1 --image "$(REPOSITORY)/$(TARGET):$(DOCKERTAG)" --arp --leaderElection --controlplane --services --inCluster > ./docs/manifests/$(VERSION)/kube-vip-arp-ds.yaml
+	@./kube-vip manifest daemonset --interface eth0 --vip 192.168.0.1 --image "$(REPOSITORY)/$(TARGET):$(DOCKERTAG)" --arp --leaderElection --controlplane --services --inCluster --enableLoadBalancer > ./docs/manifests/$(VERSION)/kube-vip-arp-ds-lb.yaml
+	@./kube-vip manifest daemonset --interface eth0 --vip 192.168.0.1 --image "$(REPOSITORY)/$(TARGET):$(DOCKERTAG)" --bgp --leaderElection --controlplane --services --inCluster > ./docs/manifests/$(VERSION)/kube-vip-bgp-ds.yaml
+	@./kube-vip manifest daemonset --interface eth0 --vip 192.168.0.1 --image "$(REPOSITORY)/$(TARGET):$(DOCKERTAG)" --bgp --leaderElection --controlplane --services --inCluster > ./docs/manifests/$(VERSION)/kube-vip-bgp-em-ds.yaml
 	@-rm ./kube-vip
 
 manifest-test: 
-	docker run $(REPOSITORY)/$(TARGET):$(DOCKERTAG) manifest pod --interface eth0 --vip 192.168.0.1 --arp --leaderElection --controlplane --services
-	docker run $(REPOSITORY)/$(TARGET):$(DOCKERTAG) manifest pod --interface eth0 --vip 192.168.0.1 --arp --leaderElection --controlplane --services --enableLoadBalancer
-	docker run $(REPOSITORY)/$(TARGET):$(DOCKERTAG) manifest pod --interface eth0 --vip 192.168.0.1 --bgp --controlplane --services 
-	docker run $(REPOSITORY)/$(TARGET):$(DOCKERTAG) manifest daemonset --interface eth0 --vip 192.168.0.1 --arp --leaderElection --controlplane --services --inCluster
-	docker run $(REPOSITORY)/$(TARGET):$(DOCKERTAG) manifest daemonset --interface eth0 --vip 192.168.0.1 --arp --leaderElection --controlplane --services --inCluster --enableLoadBalancer
-	docker run $(REPOSITORY)/$(TARGET):$(DOCKERTAG) manifest daemonset --interface eth0 --vip 192.168.0.1 --bgp --leaderElection --controlplane --services --inCluster
+	docker run $(REPOSITORY)/$(TARGET):$(DOCKERTAG) manifest pod --interface eth0 --vip 192.168.0.1 --image "$(REPOSITORY)/$(TARGET):$(DOCKERTAG)" --arp --leaderElection --controlplane --services
+	docker run $(REPOSITORY)/$(TARGET):$(DOCKERTAG) manifest pod --interface eth0 --vip 192.168.0.1 --image "$(REPOSITORY)/$(TARGET):$(DOCKERTAG)" --arp --leaderElection --controlplane --services --enableLoadBalancer
+	docker run $(REPOSITORY)/$(TARGET):$(DOCKERTAG) manifest pod --interface eth0 --vip 192.168.0.1 --image "$(REPOSITORY)/$(TARGET):$(DOCKERTAG)" --bgp --controlplane --services
+	docker run $(REPOSITORY)/$(TARGET):$(DOCKERTAG) manifest daemonset --interface eth0 --vip 192.168.0.1 --image "$(REPOSITORY)/$(TARGET):$(DOCKERTAG)" --arp --leaderElection --controlplane --services --inCluster
+	docker run $(REPOSITORY)/$(TARGET):$(DOCKERTAG) manifest daemonset --interface eth0 --vip 192.168.0.1 --image "$(REPOSITORY)/$(TARGET):$(DOCKERTAG)" --arp --leaderElection --controlplane --services --inCluster --enableLoadBalancer
+	docker run $(REPOSITORY)/$(TARGET):$(DOCKERTAG) manifest daemonset --interface eth0 --vip 192.168.0.1 --image "$(REPOSITORY)/$(TARGET):$(DOCKERTAG)" --bgp --leaderElection --controlplane --services --inCluster
 
 unit-tests:
-	go test ./...
+	go test -race ./...
+
+unit-tests-docker:
+	docker run --rm -w /kube-vip -v $$(pwd):/kube-vip -v kube-vip-gomod-cache:/go/pkg/mod -v kube-vip-gobuild-cache:/root/.cache/go-build golang:$(GO_VERSION) make unit-tests
 
 integration-tests:
 	go test -tags=integration,e2e -v ./pkg/etcd
 
-e2e-tests:
-	GOMAXPROCS=4 E2E_IMAGE_PATH=$(REPOSITORY)/$(TARGET):$(DOCKERTAG) go run github.com/onsi/ginkgo/v2/ginkgo --tags=e2e -v -p ./testing/e2e ./testing/e2e/etcd
+e2e-tests-arp: get-whoami
+	GOMAXPROCS=4 TEST_MODE=arp K8S_IMAGE_PATH=kindest/node:$(K8S_VERSION) E2E_IMAGE_PATH=$(REPOSITORY)/$(TARGET):$(DOCKERTAG) go run github.com/onsi/ginkgo/v2/ginkgo --tags=e2e -v -p ./testing/e2e
 
-e2e-tests129-arp:
-	GOMAXPROCS=4 TEST_MODE=arp V129=true K8S_IMAGE_PATH=kindest/node:v1.29.0 E2E_IMAGE_PATH=$(REPOSITORY)/$(TARGET):$(DOCKERTAG) go run github.com/onsi/ginkgo/v2/ginkgo --tags=e2e -v -p ./testing/e2e
+e2e-tests-rt: get-whoami
+	GOMAXPROCS=4 TEST_MODE=rt K8S_IMAGE_PATH=kindest/node:$(K8S_VERSION) E2E_IMAGE_PATH=$(REPOSITORY)/$(TARGET):$(DOCKERTAG) go run github.com/onsi/ginkgo/v2/ginkgo --tags=e2e -v -p ./testing/e2e
 
-e2e-tests129-rt:
-	GOMAXPROCS=4 TEST_MODE=rt V129=true K8S_IMAGE_PATH=kindest/node:v1.29.0 E2E_IMAGE_PATH=$(REPOSITORY)/$(TARGET):$(DOCKERTAG) go run github.com/onsi/ginkgo/v2/ginkgo --tags=e2e -v -p ./testing/e2e
+e2e-tests-bgp: get-whoami get-gobgp
+	GOMAXPROCS=4 TEST_MODE=bgp K8S_IMAGE_PATH=kindest/node:$(K8S_VERSION) E2E_IMAGE_PATH=$(REPOSITORY)/$(TARGET):$(DOCKERTAG) go run github.com/onsi/ginkgo/v2/ginkgo --tags=e2e -v -p ./testing/e2e
 
-e2e-tests129-bgp:
-	GOMAXPROCS=4 TEST_MODE=bgp V129=true K8S_IMAGE_PATH=kindest/node:v1.29.0 E2E_IMAGE_PATH=$(REPOSITORY)/$(TARGET):$(DOCKERTAG) go run github.com/onsi/ginkgo/v2/ginkgo --tags=e2e -v -p ./testing/e2e
-
-e2e-tests129: e2e-tests129-arp e2e-tests129-rt e2e-tests129-bgp
+e2e-tests: e2e-tests-arp e2e-tests-rt e2e-tests-bgp
 
 service-tests:
-	E2E_IMAGE_PATH=$(REPOSITORY)/$(TARGET):$(DOCKERTAG) go run ./testing/services -Services -simple -deployments -leaderActive -leaderFailover -localDeploy -egress -egressIPv6 -dualStack
+	$(MAKE) -C testing/e2e/e2e dockerLocal
+	E2E_IMAGE_PATH=$(REPOSITORY)/$(TARGET):$(DOCKERTAG) go run ./testing/services -Services -simple -deployments -leaderActive -leaderFailover -localDeploy -electionFaults -egress -egressIPv6 -dualStack -egressInternal
 
 trivy: dockerx86ActionIPTables
 	docker run -v /var/run/docker.sock:/var/run/docker.sock aquasec/trivy:0.47.0 \
@@ -166,14 +169,16 @@ kind-quick:
 	kubectl create configmap --namespace kube-system kubevip --from-literal range-global=172.18.100.10-172.18.100.30
 	kubectl apply -f https://raw.githubusercontent.com/kube-vip/kube-vip-cloud-provider/main/manifest/kube-vip-cloud-controller.yaml
 	kind load docker-image --name kube-vip $(REPOSITORY)/$(TARGET):$(DOCKERTAG)
-	docker run --network host --rm $(REPOSITORY)/$(TARGET):$(DOCKERTAG) manifest daemonset --services --inCluster --arp --servicesElection --interface  eth0 | kubectl apply -f -
+	docker run --network host --rm $(REPOSITORY)/$(TARGET):$(DOCKERTAG) manifest daemonset --services --inCluster --image "$(REPOSITORY)/$(TARGET):$(DOCKERTAG)" --arp --servicesElection --interface  eth0 | kubectl apply -f -
 
 kind-reload:
-	kind load docker-image $(REPOSITORY)/$(TARGET):$(DOCKERTAG) --name kube-vip
+	kind load docker-image $(REPOSITORY)/$(TARGET):$(DOCKERTAG) --name services
 	kubectl rollout restart -n kube-system daemonset/kube-vip-ds
 
 get-gobgp:
 	mkdir -p bin
-	wget -nc --directory-prefix=bin https://github.com/osrg/gobgp/releases/download/v3.37.0/gobgp_3.37.0_linux_amd64.tar.gz
-	tar -xvzf bin/gobgp_3.37.0_linux_amd64.tar.gz -C bin
+	wget -nc --directory-prefix=bin https://github.com/osrg/gobgp/releases/download/v4.6.0/gobgp_4.6.0_linux_amd64.tar.gz
+	tar -xvzf bin/gobgp_4.6.0_linux_amd64.tar.gz -C bin
 
+get-whoami:
+	docker pull ghcr.io/traefik/whoami:v1.11

README.md

@@ -18,6 +18,8 @@ The idea behind `kube-vip` is a small self-contained Highly-Available option for
 
 **NOTE:** All documentation of both usage and architecture are now available at [https://kube-vip.io](https://kube-vip.io).
 
+For upgrading an existing install in place (static Pod or DaemonSet), see the [upgrade guide](https://kube-vip.io/docs/upgrade/).
+
 ## Features
 
 Kube-Vip was originally created to provide a HA solution for the Kubernetes control plane, over time it has evolved to incorporate that same functionality into Kubernetes service type [load-balancers](https://kubernetes.io/docs/concepts/services-networking/service/#loadbalancer).
@@ -58,14 +60,81 @@ All of these would require a separate level of configuration and in some infrast
 
 ## Troubleshooting and Feedback
 
+### SELinux and IPVS kernel modules
+
+When using IPVS load balancing on nodes with SELinux enforcing, kube-vip may be
+blocked from requesting kernel modules from inside the container. Symptoms can
+include the kube-vip pod entering `Error` or `CrashLoopBackOff`, logs that show
+`ensure IPVS kernel modules are loaded`, or audit denials for `module_request`
+from `container_t`.
+
+Load the required IPVS modules on every node that can run kube-vip before
+deploying it:
+
+```shell
+sudo modprobe ip_vs
+sudo modprobe ip_vs_rr
+```
+
+To persist this across reboots, add the modules to a file such as
+`/etc/modules-load.d/kube-vip-ipvs.conf`:
+
+```text
+ip_vs
+ip_vs_rr
+```
+
+Preloading only the required modules is preferred to enabling the SELinux
+`domain_kernel_load_modules` boolean for containers.
+
+### Gateway API `LoadBalancer` services with no endpoints
+
+Some Gateway API controllers create `LoadBalancer` services that intentionally have no Endpoints/EndpointSlices backends.
+
+If you want kube-vip to reconcile such a service, opt in with:
+
+```yaml
+metadata:
+  annotations:
+    kube-vip.io/allow-reconcile-without-endpoints: "true"
+spec:
+  type: LoadBalancer
+  externalTrafficPolicy: Cluster
+```
+
+Scope:
+- Works only with `externalTrafficPolicy: Cluster`
+- No effect for `Local`
+- Default endpoint-gated behavior remains unchanged for services without this annotation
+
 Please raise issues on the GitHub repository and as mentioned check the documentation at [https://kube-vip.io](https://kube-vip.io/).
 
+## Community Tools
+
+- **[KubeStellar Console — Guided kube-vip Install](https://console.kubestellar.io/missions/install-kube-vip)** — A step-by-step guided installation experience for kube-vip with pre-flight checks, validation, troubleshooting, and rollback support.
+
 ## Contributing
 
 Thanks for taking the time to join our community and start contributing! We welcome pull requests. Feel free to dig through the [issues](https://github.com/kube-vip/kube-vip/issues) and jump in.
 
 :warning: This project has issue compiling on MacOS, please compile it on linux distribution
 
+Additionally it is now relatively easy and quick to develop with [skaffold](https://skaffold.dev/), and the `skaffold.yaml` exists within the root folder of the gir repository.
+
+### Set up a kind development environment
+
+1. `kind create cluster --config ./testing/kind.yaml`
+2. `kubectl apply -f https://kube-vip.io/manifests/rbac.yaml`
+3. Create a load balancer range `configMap` from the kind cluster
+4. Apply the CCM manifest
+5. Start skaffold with `skaffold dev`
+6. Start developing and see changes applied in real-time.
+
 ## Star History
 
 [![Star History Chart](https://api.star-history.com/svg?repos=kube-vip/kube-vip&type=Date)](https://star-history.com/#kube-vip/kube-vip&Date)
+[![FOSSA Status](https://app.fossa.com/api/projects/git%2Bgithub.com%2Fkube-vip%2Fkube-vip.svg?type=shield)](https://app.fossa.com/projects/git%2Bgithub.com%2Fkube-vip%2Fkube-vip?ref=badge_shield)
+
+
+## License
+[![FOSSA Status](https://app.fossa.com/api/projects/git%2Bgithub.com%2Fkube-vip%2Fkube-vip.svg?type=large)](https://app.fossa.com/projects/git%2Bgithub.com%2Fkube-vip%2Fkube-vip?ref=badge_large)

cmd/kube-vip-kubeadm.go

@@ -23,25 +23,45 @@ func init() {
 var kubeKubeadm = &cobra.Command{
 	Use:   "kubeadm",
 	Short: "Kubeadm functions",
-	Run: func(cmd *cobra.Command, args []string) { //nolint TODO
+	Long: `This command group provides utilities for generating static Pod manifests specifically tailored for the kubeadm bootstrapping process.
+	It contains two subcommands:
+  		- init:  Generates a manifest to be used during 'kubeadm init' on the first control-plane node.
+  		- join:  Generates a manifest to be used during 'kubeadm join' for additional control-plane nodes.
+
+The generated YAML manifest should be saved to the kubeadm static Pod directory (typically /etc/kubernetes/manifests/) so that kubeadm launches the kube-vip static Pod automatically.`,
+	Run: func(cmd *cobra.Command, _ []string) {
 		_ = cmd.Help()
-		// TODO - A load of text detailing what's actually happening
 	},
 }
 
 var kubeKubeadmInit = &cobra.Command{
 	Use:   "init",
 	Short: "kube-vip init",
-	Long:  "The \"init\" subcommand will generate the Kubernetes manifest that will be started by kubeadm through the kubeadm init process",
-	Run: func(cmd *cobra.Command, args []string) { //nolint TODO
+	Long: `The 'init' subcommand generates a Kubernetes Pod manifest that kubeadm will start as a static Pod during the cluster initialisation phase.
+
+This manifest runs kube-vip on the first control-plane node to advertise the Virtual IP (VIP) for the API server. The VIP is typically configured using ARP (Layer 2) or BGP (dynamic routing).
+
+Required flags for this command:
+  --interface   : The network interface to bind the VIP to (e.g., eth0).
+  --vip or --address : The Virtual IP address or DNS name to use.
+
+Example:
+  kube-vip kubeadm init --interface eth0 --vip 192.168.1.100 --controlplane
+
+The output YAML should be written to the kubeadm manifests directory, e.g.:
+  kube-vip kubeadm init ... > /etc/kubernetes/manifests/kube-vip.yaml`,
+	Run: func(cmd *cobra.Command, _ []string) {
 
 		initConfig.LoadBalancers = append(initConfig.LoadBalancers, initLoadBalancer)
-		// TODO - A load of text detailing what's actually happening
 		err := kubevip.ParseEnvironment(&initConfig)
 		if err != nil {
 			log.Error("parsing environment", "err", err)
 			return
 		}
+		if err := initConfig.Validate(); err != nil {
+			log.Error("validating configuration", "err", err)
+			return
+		}
 
 		// TODO - check for certain things VIP/interfaces
 		if initConfig.Interface == "" {
@@ -58,14 +78,18 @@ var kubeKubeadmInit = &cobra.Command{
 
 		// Ensure there is an address to generate the CIDR from
 		if initConfig.VIPSubnet == "" && initConfig.Address != "" {
-			initConfig.VIPSubnet, err = GenerateCidrRange(initConfig.Address)
+			initConfig.VIPSubnet, err = GenerateCidrRange(initConfig.Address, initConfig.DNSMode)
 			if err != nil {
 				log.Error("generating VIPSubnet", "err", err)
 				return
 			}
 		}
 
-		cfg := kubevip.GeneratePodManifestFromConfig(&initConfig, Release.Version, inCluster)
+		cfg, err := kubevip.GeneratePodManifestFromConfig(&initConfig, image, Release.Version, inCluster)
+		if err != nil {
+			log.Error("unable to create manifest", "err", err)
+			return
+		}
 		fmt.Println(cfg) // output manifest to stdout
 	},
 }
@@ -73,15 +97,30 @@ var kubeKubeadmInit = &cobra.Command{
 var kubeKubeadmJoin = &cobra.Command{
 	Use:   "join",
 	Short: "kube-vip join",
-	Run: func(cmd *cobra.Command, args []string) { //nolint TODO
+	Long: `The 'join' subcommand generates a Kubernetes Pod manifest for additional control-plane nodes joining an existing cluster via 'kubeadm join'.
+
+It functions identically to the 'init' subcommand, but is intended for secondary control-plane nodes. It validates that the kubeconfig file (specified by --config, defaulting to /etc/kubernetes/admin.conf) exists on the node to ensure the node can authenticate with the cluster.
+
+Required flags for this command:
+  --interface   : The network interface to bind the VIP to.
+  --vip or --address : The Virtual IP address or DNS name (must match the VIP used during 'init').
+
+Example:
+  kube-vip kubeadm join --interface eth0 --vip 192.168.1.100
+
+The output YAML should be saved to the kubeadm manifests directory on the joining node.`,
+	Run: func(cmd *cobra.Command, _ []string) {
 
 		initConfig.LoadBalancers = append(initConfig.LoadBalancers, initLoadBalancer)
-		// TODO - A load of text detailing what's actually happening
 		err := kubevip.ParseEnvironment(&initConfig)
 		if err != nil {
 			log.Error("parsing environment", "err", err)
 			return
 		}
+		if err := initConfig.Validate(); err != nil {
+			log.Error("validating configuration", "err", err)
+			return
+		}
 
 		// TODO - check for certain things VIP/interfaces
 		if initConfig.Interface == "" {
@@ -103,14 +142,18 @@ var kubeKubeadmJoin = &cobra.Command{
 
 		// Ensure there is an address to generate the CIDR from
 		if initConfig.VIPSubnet == "" && initConfig.Address != "" {
-			initConfig.VIPSubnet, err = GenerateCidrRange(initConfig.Address)
+			initConfig.VIPSubnet, err = GenerateCidrRange(initConfig.Address, initConfig.DNSMode)
 			if err != nil {
 				log.Error("generating VIPSubnet", "err", err)
 				return
 			}
 		}
 
-		cfg := kubevip.GeneratePodManifestFromConfig(&initConfig, Release.Version, inCluster)
+		cfg, err := kubevip.GeneratePodManifestFromConfig(&initConfig, image, Release.Version, inCluster)
+		if err != nil {
+			log.Error("unable to create manifest", "err", err)
+			return
+		}
 		fmt.Println(cfg) // output manifest to stdout
 	},
 }

cmd/kube-vip-manifests.go

@@ -19,6 +19,7 @@ var taint, role, rolebinding bool
 
 func init() {
 	kubeManifest.PersistentFlags().BoolVar(&inCluster, "inCluster", false, "Use the incluster token to authenticate to Kubernetes")
+	kubeManifest.PersistentFlags().StringVar(&image, "image", "ghcr.io/kube-vip/kube-vip", "Define a hardcoded image with or without tag for the manifest")
 	kubeManifestDaemon.PersistentFlags().BoolVar(&taint, "taint", false, "Taint the manifest for only running on control planes")
 	kubeManifestRbac.PersistentFlags().BoolVar(&role, "role", false, "Generate only a Role inside the serviceNamespace access")
 	kubeManifestRbac.PersistentFlags().BoolVar(&rolebinding, "rolebinding", false, "Generate only a RoleBinding for namespaced access")
@@ -31,42 +32,71 @@ func init() {
 var kubeManifest = &cobra.Command{
 	Use:   "manifest",
 	Short: "Manifest functions",
-	Run: func(cmd *cobra.Command, args []string) { //nolint TODO
+	Long: `This command group provides flexible manifest generation for deploying kube-vip in various Kubernetes environments.
+
+Unlike the 'kubeadm' subcommands, which are tightly coupled to kubeadm's static Pod requirements, these generators produce standard Kubernetes manifests (Pod, DaemonSet, RBAC) that can be used with any Kubernetes distribution (e.g., k3s, RKE, or vanilla Kubernetes).
+
+Subcommands:
+  pod        : Generates a standalone Pod manifest (similar to a static pod).
+  daemonset  : Generates a DaemonSet manifest to run kube-vip on selected nodes.
+  rbac       : Generates the necessary ServiceAccount, Role/ClusterRole, and Binding manifests.
+
+All output is written to stdout as YAML, typically piped to 'kubectl apply -f -' or saved to a file.`,
+	Run: func(cmd *cobra.Command, _ []string) {
 		_ = cmd.Help()
-		// TODO - A load of text detailing what's actually happening
 	},
 }
 
 var kubeManifestPod = &cobra.Command{
 	Use:   "pod",
 	Short: "Generate a Pod Manifest",
-	Run: func(cmd *cobra.Command, args []string) { //nolint TODO
+	Long: `Generate a standalone Pod manifest for kube-vip.
+
+This is ideal for environments that do not use DaemonSets or where you want to run kube-vip as a static Pod (similar to the 'kubeadm' subcommand, but without kubeadm-specific assumptions). It includes all the necessary container specifications, volumes, and environment variables derived from the provided flags.
+
+Key flags:
+  --interface : Network interface for the VIP.
+  --vip or --address : The Virtual IP address or DNS name.
+  --image     : Override the container image (default: ghcr.io/kube-vip/kube-vip).
+
+The manifest is generated based on the current configuration flags set on the root command.
+
+Example:
+  kube-vip manifest pod --interface eth0 --vip 10.0.0.100 --controlplane | kubectl apply -f -`,
+	Run: func(cmd *cobra.Command, _ []string) {
 		var err error
 
 		initConfig.LoadBalancers = append(initConfig.LoadBalancers, initLoadBalancer)
-		// TODO - A load of text detailing what's actually happening
 		if err := kubevip.ParseEnvironment(&initConfig); err != nil {
 			log.Error("parsing environment", "err", err)
 			return
 		}
+		if err := initConfig.Validate(); err != nil {
+			log.Error("validating configuration", "err", err)
+			return
+		}
 
 		// The control plane has a requirement for a VIP being specified
 		if initConfig.EnableControlPlane && (initConfig.VIP == "" && initConfig.Address == "" && !initConfig.DDNS) {
 			_ = cmd.Help()
-			log.Error("No address is specified for kube-vip to expose services on")
+			log.Error("no address is specified for kube-vip to expose services on")
 			return
 		}
 
 		// Ensure there is an address to generate the CIDR from
 		if initConfig.VIPSubnet == "" && initConfig.Address != "" {
-			initConfig.VIPSubnet, err = GenerateCidrRange(initConfig.Address)
+			initConfig.VIPSubnet, err = GenerateCidrRange(initConfig.Address, initConfig.DNSMode)
 			if err != nil {
 				log.Error("config parse", "err", err)
 				return
 			}
 		}
 
-		cfg := kubevip.GeneratePodManifestFromConfig(&initConfig, Release.Version, inCluster)
+		cfg, err := kubevip.GeneratePodManifestFromConfig(&initConfig, image, Release.Version, inCluster)
+		if err != nil {
+			log.Error("unable to create manifest", "err", err)
+			return
+		}
 		fmt.Println(cfg) // output manifest to stdout
 	},
 }
@@ -74,32 +104,50 @@ var kubeManifestPod = &cobra.Command{
 var kubeManifestDaemon = &cobra.Command{
 	Use:   "daemonset",
 	Short: "Generate a Daemonset Manifest",
-	Run: func(cmd *cobra.Command, args []string) { //nolint TODO
+	Long: `Generate a DaemonSet manifest to run kube-vip across multiple nodes.
+
+This is the recommended deployment method for production clusters running kube-vip as a service. It ensures that kube-vip runs on all control-plane nodes (or selected nodes via tolerations) and can handle both control-plane HA and service load-balancing.
+
+Flags specific to this subcommand:
+  --taint  : Adds a toleration to the DaemonSet so that pods are scheduled only on nodes with the control-plane taint (node-role.kubernetes.io/control-plane:NoSchedule). This is essential for control-plane-only deployments.
+
+All other standard kube-vip flags (--interface, --vip, --enableARP, --enableBGP, etc.) are respected and embedded into the DaemonSet pod template.
+
+Example:
+  kube-vip manifest daemonset --interface eth0 --vip 192.168.1.100 --controlplane --taint | kubectl apply -f -`,
+	Run: func(cmd *cobra.Command, _ []string) {
 		var err error
 
 		initConfig.LoadBalancers = append(initConfig.LoadBalancers, initLoadBalancer)
-		// TODO - A load of text detailing what's actually happening
 		if err := kubevip.ParseEnvironment(&initConfig); err != nil {
 			log.Error("parsing environment", "err", err)
 			return
 		}
+		if err := initConfig.Validate(); err != nil {
+			log.Error("validating configuration", "err", err)
+			return
+		}
 		// The control plane has a requirement for a VIP being specified
 		if initConfig.EnableControlPlane && (initConfig.VIP == "" && initConfig.Address == "" && !initConfig.DDNS) {
 			_ = cmd.Help()
-			log.Error("No address is specified for kube-vip to expose services on")
+			log.Error("no address is specified for kube-vip to expose services on")
 			return
 		}
 
 		// Ensure there is an address to generate the CIDR from
 		if initConfig.VIPSubnet == "" && initConfig.Address != "" {
-			initConfig.VIPSubnet, err = GenerateCidrRange(initConfig.Address)
+			initConfig.VIPSubnet, err = GenerateCidrRange(initConfig.Address, initConfig.DNSMode)
 			if err != nil {
 				log.Error("config parse", "err", err)
 				return
 			}
 		}
 
-		cfg := kubevip.GenerateDaemonsetManifestFromConfig(&initConfig, Release.Version, inCluster, taint)
+		cfg, err := kubevip.GenerateDaemonsetManifestFromConfig(&initConfig, image, Release.Version, inCluster, taint)
+		if err != nil {
+			log.Error("unable to create manifest", "err", err)
+			return
+		}
 		fmt.Println(cfg) // output manifest to stdout
 	},
 }
@@ -107,26 +155,40 @@ var kubeManifestDaemon = &cobra.Command{
 var kubeManifestRbac = &cobra.Command{
 	Use:   "rbac",
 	Short: "Generate an RBAC Manifest",
-	Run: func(cmd *cobra.Command, args []string) { //nolint TODO
-		var err error
+	Long: `Generate the RBAC (Role-Based Access Control) manifests required for kube-vip to interact with the Kubernetes API.
 
+kube-vip needs permissions to watch services, endpoints, configmaps, and manage leader election leases. This command outputs the minimum required ServiceAccount, Role (or ClusterRole), and the corresponding binding.
+
+Flags:
+  --role        : If true, generates a namespaced Role instead of a ClusterRole. The namespace is taken from the root --namespace flag (default: kube-system).
+  --rolebinding : If true, generates a RoleBinding (if --role is also true). If --role is false, a ClusterRoleBinding is generated automatically.
+
+The output is a multi-document YAML (separated by '---'). It is safe to apply directly:
+  kube-vip manifest rbac --role --rolebinding | kubectl apply -f -
+
+Without --role, it generates a ClusterRole and ClusterRoleBinding, which is the default behaviour and suitable for most cluster-wide deployments.`,
+	Run: func(cmd *cobra.Command, _ []string) {
 		initConfig.LoadBalancers = append(initConfig.LoadBalancers, initLoadBalancer)
-		// TODO - A load of text detailing what's actually happening
 		if err := kubevip.ParseEnvironment(&initConfig); err != nil {
 			log.Error("parsing environment", "err", err)
 			return
 		}
+		if err := initConfig.Validate(); err != nil {
+			log.Error("validating configuration", "err", err)
+			return
+		}
 
 		// The control plane has a requirement for a VIP being specified
 		if initConfig.EnableControlPlane && (initConfig.VIP == "" && initConfig.Address == "" && !initConfig.DDNS) {
 			_ = cmd.Help()
-			log.Error("No address is specified for kube-vip to expose services on")
+			log.Error("no address is specified for kube-vip to expose services on")
 			return
 		}
 
 		// Ensure there is an address to generate the CIDR from
 		if initConfig.VIPSubnet == "" && initConfig.Address != "" {
-			initConfig.VIPSubnet, err = GenerateCidrRange(initConfig.Address)
+			var err error
+			initConfig.VIPSubnet, err = GenerateCidrRange(initConfig.Address, initConfig.DNSMode)
 			if err != nil {
 				log.Error("generating VIPSubnet", "err", err)
 				return

cmd/kube-vip.go

@@ -6,22 +6,30 @@ import (
 	"net"
 	"net/http"
 	"os"
+	"slices"
+	"strconv"
 	"strings"
+	"sync"
 	"time"
 
 	log "log/slog"
 
-	"github.com/prometheus/client_golang/prometheus"
 	"github.com/prometheus/client_golang/prometheus/promhttp"
 	"github.com/spf13/cobra"
 	"github.com/vishvananda/netlink"
 	"golang.org/x/sys/unix"
 
+	"github.com/kube-vip/kube-vip/pkg/debouncer"
 	"github.com/kube-vip/kube-vip/pkg/kubevip"
 	"github.com/kube-vip/kube-vip/pkg/manager"
+	"github.com/kube-vip/kube-vip/pkg/metrics"
+	"github.com/kube-vip/kube-vip/pkg/utils"
 	"github.com/kube-vip/kube-vip/pkg/vip"
 )
 
+// Is an option to set the image
+var image string
+
 // Is kube-vip running within cluster
 var inCluster bool
 
@@ -44,14 +52,16 @@ var (
 )
 
 var kubeVipCmd = &cobra.Command{
-	Use:   "kube-vip",
-	Short: "This is a server for providing a Virtual IP and load-balancer for the Kubernetes control-plane",
+	Use:           "kube-vip",
+	Short:         "This is a server for providing a Virtual IP and load-balancer for the Kubernetes control-plane",
+	SilenceErrors: true,
 }
 
 func init() {
 	// Basic flags
 	kubeVipCmd.PersistentFlags().StringVar(&initConfig.Interface, "interface", "", "Name of the interface to bind to")
 	kubeVipCmd.PersistentFlags().StringVar(&initConfig.ServicesInterface, "serviceInterface", "", "Name of the interface to bind to (for services)")
+	kubeVipCmd.PersistentFlags().BoolVar(&initConfig.AllowInterfaceNotUp, "allowInterfaceNotUp", false, "Allow kube-vip to start even if the interface is not up")
 	kubeVipCmd.PersistentFlags().StringVar(&initConfig.VIP, "vip", "", "The Virtual IP address")
 	kubeVipCmd.PersistentFlags().StringVar(&initConfig.VIPSubnet, "vipSubnet", "", "The Virtual IP address subnet e.g. /32 /24 /8 etc.. (Default to 32 for IPv4 and 128 for IPv6)")
 	kubeVipCmd.PersistentFlags().StringVar(&initConfig.NodeName, "nodeName", "", "Name to be used for lease holder. Must be unique for each node/instance")
@@ -62,6 +72,9 @@ func init() {
 	kubeVipCmd.PersistentFlags().BoolVar(&initConfig.EnableARP, "arp", false, "Enable Arp for VIP changes")
 	kubeVipCmd.PersistentFlags().BoolVar(&initConfig.EnableWireguard, "wireguard", false, "Enable Wireguard for services VIPs")
 	kubeVipCmd.PersistentFlags().BoolVar(&initConfig.EnableRoutingTable, "table", false, "Enable Routing Table for services VIPs")
+	kubeVipCmd.PersistentFlags().BoolVar(&initConfig.PreserveVIPOnLeadershipLoss, "preserveVipOnLeadershipLoss", false, "Preserve ARP VIP addresses on interface when leadership is lost (default: false for backward compatibility)")
+	kubeVipCmd.PersistentFlags().BoolVar(&initConfig.LoseLeadership, "loseLeadership", false, "Lose leadership when VIP interface goes down")
+	kubeVipCmd.PersistentFlags().IntVar(&initConfig.LoseLeadershipTimeoutSeconds, "loseLeadershiptTimeoutSeconds", 30, "Timeout before re-electing a leader when the VIP interface is down")
 
 	// LoadBalancer flags
 	kubeVipCmd.PersistentFlags().BoolVar(&initConfig.EnableLoadBalancer, "enableLoadBalancer", false, "enable loadbalancing on the VIP with IPVS")
@@ -74,12 +87,13 @@ func init() {
 	kubeVipCmd.PersistentFlags().BoolVar(&initConfig.EnableLeaderElection, "leaderElection", false, "Use the Kubernetes leader election mechanism for clustering")
 	kubeVipCmd.PersistentFlags().StringVar(&initConfig.LeaderElectionType, "leaderElectionType", "kubernetes", "Defines the backend to run the leader election: kubernetes or etcd. Defaults to kubernetes.")
 	kubeVipCmd.PersistentFlags().StringVar(&initConfig.LeaseName, "leaseName", "plndr-cp-lock", "Name of the lease that is used for leader election")
-	kubeVipCmd.PersistentFlags().IntVar(&initConfig.LeaseDuration, "leaseDuration", 5, "Length of time (in seconds) a Kubernetes leader lease can be held for")
-	kubeVipCmd.PersistentFlags().IntVar(&initConfig.RenewDeadline, "leaseRenewDuration", 3, "Length of time (in seconds) a Kubernetes leader can attempt to renew its lease")
-	kubeVipCmd.PersistentFlags().IntVar(&initConfig.RetryPeriod, "leaseRetry", 1, "Length of time (in seconds) the LeaderElector clients should wait between tries of actions")
+	kubeVipCmd.PersistentFlags().IntVar(&initConfig.LeaseDuration, "leaseDuration", 15, "Length of time (in seconds) a Kubernetes leader lease can be held for")
+	kubeVipCmd.PersistentFlags().IntVar(&initConfig.RenewDeadline, "leaseRenewDuration", 10, "Length of time (in seconds) a Kubernetes leader can attempt to renew its lease")
+	kubeVipCmd.PersistentFlags().IntVar(&initConfig.RetryPeriod, "leaseRetry", 2, "Length of time (in seconds) the LeaderElector clients should wait between tries of actions")
 
 	// BGP flags
 	kubeVipCmd.PersistentFlags().BoolVar(&initConfig.EnableBGP, "bgp", false, "This will enable BGP support within kube-vip")
+	kubeVipCmd.PersistentFlags().BoolVar(&initConfig.BGPAttachIPToInterface, "bgpAttachIPToInterface", false, "Assign BGP service VIPs to the configured interface")
 	kubeVipCmd.PersistentFlags().StringVar(&initConfig.BGPConfig.RouterID, "bgpRouterID", "", "The routerID for the bgp server")
 	kubeVipCmd.PersistentFlags().StringVar(&initConfig.BGPConfig.SourceIF, "sourceIF", "", "The source interface for bgp peering (not to be used with sourceIP)")
 	kubeVipCmd.PersistentFlags().StringVar(&initConfig.BGPConfig.SourceIP, "sourceIP", "", "The source address for bgp peering (not to be used with sourceIF)")
@@ -97,6 +111,11 @@ func init() {
 	kubeVipCmd.PersistentFlags().StringVar(&initConfig.BGPConfig.Zebra.URL, "zebraUrl", "unix:/var/run/frr/zserv.api", "Path to the unix domain socket for connecting to Zebra daemon")
 	kubeVipCmd.PersistentFlags().Uint32Var(&initConfig.BGPConfig.Zebra.Version, "zebraVersion", 6, "Zebra API Version")
 	kubeVipCmd.PersistentFlags().StringVar(&initConfig.BGPConfig.Zebra.SoftwareName, "zebraSoftwareName", "frr8.3", "Software Name for Zebra")
+	kubeVipCmd.PersistentFlags().StringVar(&initConfig.ControlPlaneHealthCheck.Address, "controlPlaneHealthCheckAddress", "", "URL to poll for the control-plane health check when using BGP without leader election")
+	kubeVipCmd.PersistentFlags().IntVar(&initConfig.ControlPlaneHealthCheck.PeriodSeconds, "controlPlaneHealthCheckPeriodSeconds", 5, "Seconds between control-plane health checks")
+	kubeVipCmd.PersistentFlags().IntVar(&initConfig.ControlPlaneHealthCheck.TimeoutSeconds, "controlPlaneHealthCheckTimeoutSeconds", 3, "Timeout for each control-plane health check request")
+	kubeVipCmd.PersistentFlags().IntVar(&initConfig.ControlPlaneHealthCheck.FailureThreshold, "controlPlaneHealthCheckFailureThreshold", 3, "Consecutive control-plane health check failures before withdrawing the BGP route")
+	kubeVipCmd.PersistentFlags().StringVar(&initConfig.ControlPlaneHealthCheck.CAPath, "controlPlaneHealthCheckCAPath", "", "Path to CA certificate for TLS verification when the control-plane health check URL is HTTPS")
 
 	// Namespace for kube-vip
 	kubeVipCmd.PersistentFlags().StringVarP(&initConfig.Namespace, "namespace", "n", "kube-system", "The namespace for the configmap defined within the cluster")
@@ -120,17 +139,22 @@ func init() {
 
 	// Extended behaviour flags
 	kubeVipCmd.PersistentFlags().BoolVar(&initConfig.EnableServicesElection, "servicesElection", false, "Enable leader election per kubernetes service")
-	kubeVipCmd.PersistentFlags().BoolVar(&initConfig.LoadBalancerClassOnly, "lbClassOnly", false, "Enable load balancing only for services with LoadBalancerClass \"kube-vip.io/kube-vip-class\"")
-	kubeVipCmd.PersistentFlags().StringVar(&initConfig.LoadBalancerClassName, "lbClassName", "kube-vip.io/kube-vip-class", "Name of load balancer class for kube-VIP, defaults to \"kube-vip.io/kube-vip-class\"")
+	kubeVipCmd.PersistentFlags().BoolVar(&initConfig.LoadBalancerClassOnly, "lbClassOnly", false, fmt.Sprintf("Enable load balancing only for services with LoadBalancerClass %q", kubevip.LBClassName))
+	kubeVipCmd.PersistentFlags().StringVar(&initConfig.LoadBalancerClassName, "lbClassName", kubevip.LBClassName, fmt.Sprintf("Name of load balancer class for kube-VIP, defaults to %q", kubevip.LBClassName))
 	kubeVipCmd.PersistentFlags().BoolVar(&initConfig.LoadBalancerClassLegacyHandling, "lbClassNameLegacyHandling", true, "Use legacy LoadBalancer class name handling (e.g. accepting services both with empty and non-empty class)")
 	kubeVipCmd.PersistentFlags().BoolVar(&initConfig.EnableServiceSecurity, "onlyAllowTrafficServicePorts", false, "Only allow traffic to service ports, others will be dropped, defaults to false")
-	kubeVipCmd.PersistentFlags().BoolVar(&initConfig.EnableNodeLabeling, "enableNodeLabeling", false, "Enable leader node labeling with \"kube-vip.io/has-ip=<VIP address>\", defaults to false")
+	kubeVipCmd.PersistentFlags().BoolVar(&initConfig.EnableNodeLabeling, "enableNodeLabeling", false, fmt.Sprintf("Enable leader node labeling with %q, defaults to false", kubevip.HasIP))
 	kubeVipCmd.PersistentFlags().StringVar(&initConfig.ServicesLeaseName, "servicesLeaseName", "plndr-svcs-lock", "Name of the lease that is used for leader election for services (in arp mode)")
 	kubeVipCmd.PersistentFlags().StringVar(&initConfig.DNSMode, "dnsMode", "first", "Name of the mode that DNS lookup will be performed (first, ipv4, ipv6, dual)")
+	kubeVipCmd.PersistentFlags().StringVar(&initConfig.DHCPMode, "dhcpMode", "", "Mode DHCP resolving will use to obtain IP addresses (ipv4, ipv6, dual)")
+	kubeVipCmd.PersistentFlags().UintVar(&initConfig.DHCPBackoffAttempts, "dhcpBackoffAttempts", kubevip.DefaultDHCPBackoffAttempts,
+		fmt.Sprintf("number of times DHCP client will try to obtain an IP address (defaults to: %d, 0 for unlimited retries)", kubevip.DefaultDHCPBackoffAttempts))
 	kubeVipCmd.PersistentFlags().BoolVar(&initConfig.DisableServiceUpdates, "disableServiceUpdates", false, "If true, kube-vip will process services as usual, but will not update service's Status.LoadBalancer.Ingress slice")
 	kubeVipCmd.PersistentFlags().BoolVar(&initConfig.EnableEndpoints, "enableEndpoints", false, "If enabled, kube-vip will only advertise services, but will use the (deprecated since v1.33) endpoints for IP addresses")
 	kubeVipCmd.PersistentFlags().BoolVar(&initConfig.LoInterfaceGlobalScope, "loInterfaceGlobalScope", false, "If true, kube-vip will set global scope when using the lo interface, otherwise a host scope will be used by default")
 	kubeVipCmd.PersistentFlags().IntVar(&initConfig.HealthCheckPort, "healthCheckPort", 0, "If set to non-zero (> 1024), then this is the port that the healthcheck will listen on")
+	kubeVipCmd.PersistentFlags().StringVar(&initConfig.DebounceTime, "debounceTime", debouncer.DefaultTime,
+		"Configures the time that the event debouncer will wait for the events arrival (default 0s - debouncer disabled, enable with min. 200ms)")
 
 	// Prometheus HTTP Server
 	kubeVipCmd.PersistentFlags().StringVar(&initConfig.PrometheusHTTPServer, "prometheusHTTPServer", ":2112", "Host and port used to expose Prometheus metrics via an HTTP server")
@@ -144,6 +168,14 @@ func init() {
 	// Kubernetes client specific flags
 	kubeVipCmd.PersistentFlags().StringVar(&initConfig.K8sConfigFile, "k8sConfigPath", "/etc/kubernetes/admin.conf", "Path to the configuration file used with the Kubernetes client")
 
+	// Configuration file flag
+	kubeVipCmd.PersistentFlags().StringVar(&initConfig.ConfigFile, "config-file", "", "Path to a JSON/YAML configuration file to load settings from")
+	kubeVipCmd.PersistentFlags().StringVar(&initConfig.InstanceName, "instanceName", "", "Unique name for this kube-vip instance (currently used to isolate nftables egress tables)")
+
+	kubeVipCmd.PersistentFlags().BoolVar(&initConfig.EgressWithNftables, "egressWithNftables", true, "Use nftables-based egress implementation")
+
+	kubeVipCmd.PersistentFlags().BoolVar(&initConfig.PerServiceElectionOnDemand, "perServiceElectionOnDemand", false, "Allow kube-vip to use per-service election for annotated services")
+
 	kubeVipCmd.AddCommand(kubeKubeadm)
 	kubeVipCmd.AddCommand(kubeManifest)
 	kubeVipCmd.AddCommand(kubeVipManager)
@@ -156,11 +188,16 @@ func init() {
 }
 
 // Execute - starts the command parsing process
-func Execute() {
-	if err := kubeVipCmd.Execute(); err != nil {
-		fmt.Println(err)
-		os.Exit(1)
+func Execute() int {
+	cmd, err := kubeVipCmd.ExecuteC()
+	if err != nil {
+		log.Error("command failed", "err", err)
+		if cmd == kubeVipCmd {
+			_ = cmd.Usage()
+		}
+		return 1
 	}
+	return 0
 }
 
 var kubeVipVersion = &cobra.Command{
@@ -184,13 +221,24 @@ var kubeVipSample = &cobra.Command{
 var kubeVipService = &cobra.Command{
 	Use:   "service",
 	Short: "Start the Virtual IP / Load balancer as a service within a Kubernetes cluster",
-	Run: func(cmd *cobra.Command, args []string) { //nolint TODO
+	RunE: func(cmd *cobra.Command, args []string) error { //nolint TODO
+		cmd.SilenceUsage = true
 
-		// parse environment variables, these will overwrite anything loaded or flags
+		// Load configuration from file if specified (lowest priority)
+		if initConfig.ConfigFile != "" {
+			err := kubevip.MergeConfigFromFile(&initConfig, initConfig.ConfigFile)
+			if err != nil {
+				return fmt.Errorf("loading config file: %w", err)
+			}
+		}
+
+		// parse environment variables, these will overwrite anything loaded from config file
 		err := kubevip.ParseEnvironment(&initConfig)
 		if err != nil {
-			log.Error("parsing env", "err", err)
-			return
+			return fmt.Errorf("parsing environment: %w", err)
+		}
+		if err := initConfig.Validate(); err != nil {
+			return fmt.Errorf("validating configuration: %w", err)
 		}
 
 		// Change RTN_UNSPEC to default type
@@ -202,8 +250,7 @@ var kubeVipService = &cobra.Command{
 		log.SetLogLoggerLevel(log.Level(initConfig.Logging))
 
 		if err := initConfig.CheckInterface(); err != nil {
-			log.Error("checking interface", "err", err)
-			return
+			return fmt.Errorf("checking interface: %w", err)
 		}
 
 		// User Environment variables as an option to make manifest clearer
@@ -212,40 +259,61 @@ var kubeVipService = &cobra.Command{
 			configMap = envConfigMap
 		}
 
-		// Ensure there is an address to generate the CIDR from
-		if initConfig.VIPSubnet == "" && initConfig.Address != "" {
-			initConfig.VIPSubnet, err = GenerateCidrRange(initConfig.Address)
-			if err != nil {
-				log.Error("generating CIDR", "err", err)
-				return
+		// Legacy vip_address requires vip_subnet for control-plane ARP, BGP, and Routing Table modes.
+		if initConfig.EnableControlPlane &&
+			(initConfig.EnableARP || initConfig.EnableBGP || initConfig.EnableRoutingTable) {
+			if err := initConfig.CheckSubnetExists(); err != nil {
+				return fmt.Errorf("checking subnet exists if vip_address defined: %w", err)
 			}
 		}
 
+		// Ensure there is an address to generate the CIDR from
+		if initConfig.VIPSubnet == "" && initConfig.Address != "" {
+			initConfig.VIPSubnet, err = GenerateCidrRange(initConfig.Address, initConfig.DNSMode)
+			if err != nil {
+				return fmt.Errorf("generating CIDR: %w", err)
+			}
+		}
+
+		ctx, cancel := context.WithCancel(cmd.Context())
+		defer cancel()
+
 		// Define the new service manager
-		mgr, err := manager.New(configMap, &initConfig)
+		mgr, err := manager.New(ctx, configMap, &initConfig)
 		if err != nil {
-			log.Error("new manager", "err", err)
-			return
+			return fmt.Errorf("new manager: %w", err)
 		}
 
 		// Start the service manager, this will watch the config Map and construct kube-vip services for it
-		err = mgr.Start()
+		err = mgr.Start(ctx)
 		if err != nil {
-			log.Error("manager start", "err", err)
-			return
+			return fmt.Errorf("manager start: %w", err)
 		}
+		return nil
 	},
 }
 
 var kubeVipManager = &cobra.Command{
 	Use:   "manager",
 	Short: "Start the kube-vip manager",
-	Run: func(cmd *cobra.Command, args []string) { //nolint TODO
-		// parse environment variables, these will overwrite anything loaded or flags
+	RunE: func(cmd *cobra.Command, args []string) error { //nolint TODO
+		cmd.SilenceUsage = true
+
+		// Load configuration from file if specified (lowest priority)
+		if initConfig.ConfigFile != "" {
+			err := kubevip.MergeConfigFromFile(&initConfig, initConfig.ConfigFile)
+			if err != nil {
+				return fmt.Errorf("loading config file: %w", err)
+			}
+		}
+
+		// parse environment variables, these will overwrite anything loaded from config file
 		err := kubevip.ParseEnvironment(&initConfig)
 		if err != nil {
-			log.Error("parsing environment", "err", err)
-			return
+			return fmt.Errorf("parsing environment: %w", err)
+		}
+		if err := initConfig.Validate(); err != nil {
+			return fmt.Errorf("validating configuration: %w", err)
 		}
 
 		// Change RTN_UNSPEC to default type
@@ -256,41 +324,72 @@ var kubeVipManager = &cobra.Command{
 		// Set the logging level for all subsequent functions
 		log.SetLogLoggerLevel(log.Level(initConfig.Logging))
 
+		// Legacy vip_address requires vip_subnet for control-plane ARP, BGP, and Routing Table modes.
+		if initConfig.EnableControlPlane &&
+			(initConfig.EnableARP || initConfig.EnableBGP || initConfig.EnableRoutingTable) {
+			if err := initConfig.CheckSubnetExists(); err != nil {
+				return fmt.Errorf("checking subnet exists if vip_address defined: %w", err)
+			}
+		}
+
 		// Ensure there is an address to generate the CIDR from
 		if initConfig.VIPSubnet == "" && initConfig.Address != "" {
-			initConfig.VIPSubnet, err = GenerateCidrRange(initConfig.Address)
+			initConfig.VIPSubnet, err = GenerateCidrRange(initConfig.Address, initConfig.DNSMode)
 			if err != nil {
-				log.Error("No interface is specified for kube-vip to bind to")
-				return
+				return fmt.Errorf("generating CIDR: %w", err)
 			}
 		}
 
 		// Welome messages
 		log.Info("kube-vip.io", "version", Release.Version, "build", Release.Build)
 
+		wg := sync.WaitGroup{}
+		defer wg.Wait()
+
+		// create main manager context
+		ctx, cancel := context.WithCancel(cmd.Context())
+		defer cancel()
+
 		// start prometheus server
 		if initConfig.PrometheusHTTPServer != "" {
-			go servePrometheusHTTPServer(cmd.Context(), PrometheusHTTPServerConfig{
-				Addr: initConfig.PrometheusHTTPServer,
+			wg.Go(func() {
+				servePrometheusHTTPServer(ctx, PrometheusHTTPServerConfig{
+					Addr: initConfig.PrometheusHTTPServer,
+				})
 			})
 		}
 
 		// Determine the kube-vip mode
-		var mode string
+		var (
+			mode         string
+			modesEnabled int
+		)
 		if initConfig.EnableARP {
 			mode = "ARP"
+			modesEnabled++
 		}
 
 		if initConfig.EnableBGP {
 			mode = "BGP"
+			modesEnabled++
 		}
 
 		if initConfig.EnableWireguard {
 			mode = "Wireguard"
+			modesEnabled++
 		}
 
 		if initConfig.EnableRoutingTable {
 			mode = "Routing Table"
+			modesEnabled++
+		}
+
+		if mode == "" {
+			return fmt.Errorf("no valid kube-vip mode detected, ensure a supported mode is configured")
+		}
+
+		if modesEnabled > 1 {
+			return fmt.Errorf("multiple kube-vip modes detected, ensure only one mode is configured")
 		}
 
 		// Provide configuration to output/logging
@@ -298,18 +397,15 @@ var kubeVipManager = &cobra.Command{
 
 		// End if nothing is enabled
 		if !initConfig.EnableServices && !initConfig.EnableControlPlane {
-			log.Error("no features are enabled")
-			return
+			return fmt.Errorf("no features are enabled")
 		}
 
 		if !initConfig.EnableARP && strings.Contains(initConfig.VIPSubnet, kubevip.Auto) {
-			log.Error("auto subnet discovery cannot be used outside ARP mode")
-			return
+			return fmt.Errorf("auto subnet discovery cannot be used outside ARP mode")
 		}
 
 		if strings.Contains(initConfig.VIPSubnet, kubevip.Auto) && initConfig.Address != "" {
-			log.Error("auto subnet discovery cannot be used if VIP address was provided")
-			return
+			return fmt.Errorf("auto subnet discovery cannot be used if VIP address was provided")
 		}
 
 		// If we're using wireguard then all traffic goes through the wg0 interface
@@ -326,20 +422,17 @@ var kubeVipManager = &cobra.Command{
 					log.Warn("attempting to create wireguard interface", "interface not found", initConfig.Interface)
 					err = netlink.LinkAdd(&netlink.Wireguard{LinkAttrs: netlink.LinkAttrs{Name: initConfig.Interface}})
 					if err != nil {
-						log.Error("adding link", "err", err)
-						return
+						return fmt.Errorf("adding link: %w", err)
 					}
 					l, err = netlink.LinkByName(initConfig.Interface)
 					if err != nil {
-						log.Error("finding link", "err", err)
-						return
+						return fmt.Errorf("finding link: %w", err)
 					}
 				}
 			}
 			err = netlink.LinkSetUp(l)
 			if err != nil {
-				log.Error("setting link UP", "err", err)
-				return
+				return fmt.Errorf("setting link UP: %w", err)
 			}
 
 		} else { // if we're not using Wireguard then we'll need to use an actual interface
@@ -349,25 +442,23 @@ var kubeVipManager = &cobra.Command{
 				defaultIF, err := vip.GetDefaultGatewayInterface()
 				if err != nil {
 					_ = cmd.Help()
-					log.Error("detecting interface", "err", err)
-					return
+					return fmt.Errorf("detecting interface: %w", err)
 				}
 				initConfig.Interface = defaultIF.Name
 				log.Info("kube-vip bind", "interface", initConfig.Interface)
 
-				go func() {
-					if err := vip.MonitorDefaultInterface(context.TODO(), defaultIF); err != nil {
+				wg.Go(func() {
+					if err := vip.MonitorDefaultInterface(ctx, defaultIF); err != nil {
 
 						log.Error("interface monitor", "err", err)
 						return
 					}
-				}()
+				})
 			}
 		}
 		// Perform a check on the state of the interface
 		if err := initConfig.CheckInterface(); err != nil {
-			log.Error("checking interface", "err", err)
-			return
+			return fmt.Errorf("checking interface: %w", err)
 		}
 
 		// User Environment variables as an option to make manifest clearer
@@ -377,20 +468,20 @@ var kubeVipManager = &cobra.Command{
 		}
 
 		// Define the new service manager
-		mgr, err := manager.New(configMap, &initConfig)
+		mgr, err := manager.New(ctx, configMap, &initConfig)
 		if err != nil {
-			log.Error("new manager", "err", err)
-			return
+			return fmt.Errorf("new manager: %w", err)
 		}
 
-		prometheus.MustRegister(mgr.PrometheusCollector()...)
+		metrics.RegisterPrometheusMetrics()
+		metrics.BuildInfo.WithLabelValues(Release.Version, Release.Build, initConfig.NodeName)
 
 		// Start the service manager, this will watch the config Map and construct kube-vip services for it
-		err = mgr.Start()
+		err = mgr.Start(ctx)
 		if err != nil {
-			log.Error("start manager", "err", err)
-			return
+			return fmt.Errorf("start manager: %w", err)
 		}
+		return nil
 	},
 }
 
@@ -420,19 +511,20 @@ func servePrometheusHTTPServer(ctx context.Context, config PrometheusHTTPServerC
 		ReadHeaderTimeout: 2 * time.Second,
 	}
 
-	go func() {
+	wg := sync.WaitGroup{}
+
+	wg.Go(func() {
 		if err = srv.ListenAndServe(); err != nil && err != http.ErrServerClosed {
 			log.Error("prometheus HTTP server", "err", err)
 			return
 		}
-	}()
+	})
 
 	log.Info("prometheus HTTP server started")
 
 	<-ctx.Done()
 
-	log.Info("prometheus HTTP server stopped")
-
+	// create prometheus shutdown context (independent of other contexts)
 	ctxShutDown, cancel := context.WithTimeout(context.Background(), 5*time.Second)
 	defer func() {
 		cancel()
@@ -446,28 +538,43 @@ func servePrometheusHTTPServer(ctx context.Context, config PrometheusHTTPServerC
 	if err == http.ErrServerClosed {
 		err = nil
 	}
+
+	log.Info("prometheus HTTP server stopped")
+
+	wg.Wait()
 }
 
-func GenerateCidrRange(address string) (string, error) {
+func GenerateCidrRange(address string, dnsMode string) (string, error) {
 	var cidrs []string
 
 	addresses := strings.Split(address, ",")
 	for _, a := range addresses {
 		ip := net.ParseIP(a)
 		if ip == nil {
-			ips, err := net.LookupIP(a)
+			// we probably are a DNS name
+			ips, err := utils.LookupHost(a, dnsMode, true)
 			if len(ips) == 0 || err != nil {
 				return "", fmt.Errorf("invalid IP address: %s from [%s], %v", a, address, err)
 			}
-			ip = ips[0]
-		}
-
-		if ip.To4() != nil {
-			cidrs = append(cidrs, "32")
+			for _, addr := range ips {
+				ip = net.ParseIP(addr)
+				if ip.To4() != nil {
+					cidrs = append(cidrs, strconv.Itoa(vip.DefaultMaskIPv4))
+				} else {
+					cidrs = append(cidrs, strconv.Itoa(vip.DefaultMaskIPv6))
+				}
+			}
 		} else {
-			cidrs = append(cidrs, "128")
+			if ip.To4() != nil {
+				cidrs = append(cidrs, strconv.Itoa(vip.DefaultMaskIPv4))
+			} else {
+				cidrs = append(cidrs, strconv.Itoa(vip.DefaultMaskIPv6))
+			}
 		}
 	}
-
+	// compact as DNS could have a lot of addresses
+	slices.Sort(cidrs)
+	cidrs = slices.Compact(cidrs)
+	slices.Reverse(cidrs)
 	return strings.Join(cidrs, ","), nil
 }

demo/client/main.go

@@ -19,7 +19,7 @@ func main() {
 	var errorOccurred bool
 	for {
 		p := make([]byte, 2048)
-		conn, err := net.Dial("udp", fmt.Sprintf("%s:%d", *address, *port))
+		conn, err := net.Dial("udp", net.JoinHostPort(*address, fmt.Sprint(port)))
 		if err != nil {
 			if !errorOccurred {
 				errorTime = time.Now()

go.mod

@@ -1,165 +1,154 @@
 module github.com/kube-vip/kube-vip
 
-go 1.24.4
+go 1.26.4
 
 require (
-	github.com/cloudflare/ipvs v0.11.0
-	github.com/davecgh/go-spew v1.1.2-0.20180830191138-d8f796af33cc
-	github.com/docker/docker v28.3.3+incompatible
-	github.com/florianl/go-conntrack v0.4.0
+	github.com/cloudflare/ipvs v0.12.0
+	github.com/containernetworking/plugins v1.9.1
+	github.com/docker/docker v28.5.2+incompatible
+	github.com/florianl/go-conntrack v0.7.0
 	github.com/google/go-cmp v0.7.0
+	github.com/google/go-containerregistry v0.21.9
 	github.com/google/nftables v0.3.0
-	github.com/gookit/slog v0.5.8
+	github.com/gookit/slog v0.7.1
 	github.com/huin/goupnp v1.3.0
-	github.com/insomniacslk/dhcp v0.0.0-20241224095048-b56fa0d5f25d
+	github.com/insomniacslk/dhcp v0.0.0-20260719225207-c76316d4aa82
 	github.com/jpillora/backoff v1.0.0
 	github.com/mdlayher/ndp v1.1.0
-	github.com/onsi/ginkgo/v2 v2.23.4
-	github.com/onsi/gomega v1.37.0
-	github.com/osrg/gobgp/v3 v3.37.0
+	github.com/onsi/ginkgo/v2 v2.32.1
+	github.com/onsi/gomega v1.42.1
+	github.com/osrg/gobgp/v4 v4.8.0
 	github.com/pkg/errors v0.9.1
-	github.com/prometheus/client_golang v1.22.0
-	github.com/sirupsen/logrus v1.9.3
-	github.com/spf13/cobra v1.9.1
-	github.com/stretchr/testify v1.10.0
+	github.com/prometheus/client_golang v1.24.1
+	github.com/spf13/cobra v1.10.2
+	github.com/stretchr/testify v1.11.1
 	github.com/vishvananda/netlink v1.3.1
-	go.etcd.io/etcd/api/v3 v3.6.1
-	go.etcd.io/etcd/client/pkg/v3 v3.6.1
-	go.etcd.io/etcd/client/v3 v3.6.1
-	go.uber.org/zap v1.27.0
+	go.etcd.io/etcd/api/v3 v3.7.1
+	go.etcd.io/etcd/client/pkg/v3 v3.7.1
+	go.etcd.io/etcd/client/v3 v3.7.1
+	go.uber.org/zap v1.28.0
 	golang.org/x/exp v0.0.0-20250103183323-7d7fa50e5329
-	golang.org/x/sync v0.15.0
-	golang.org/x/sys v0.33.0
+	golang.org/x/sync v0.22.0
+	golang.org/x/sys v0.47.0
 	golang.zx2c4.com/wireguard/wgctrl v0.0.0-20241231184526-a9ab2273dd10
-	google.golang.org/grpc v1.73.0
-	google.golang.org/protobuf v1.36.6
-	k8s.io/api v0.33.3
-	k8s.io/apimachinery v0.33.3
-	k8s.io/client-go v0.33.3
-	k8s.io/klog/v2 v2.130.1
-	sigs.k8s.io/kind v0.29.0
-	sigs.k8s.io/yaml v1.4.0
+	google.golang.org/grpc v1.83.0
+	gopkg.in/yaml.v3 v3.0.1
+	k8s.io/api v0.36.3
+	k8s.io/apimachinery v0.36.3
+	k8s.io/client-go v0.36.3
+	k8s.io/klog/v2 v2.140.0
+	sigs.k8s.io/kind v0.32.0
+	sigs.k8s.io/yaml v1.6.0
 )
 
 require (
 	al.essio.dev/pkg/shellescape v1.5.1 // indirect
-	github.com/BurntSushi/toml v1.4.0 // indirect
-	github.com/Microsoft/go-winio v0.4.14 // indirect
+	github.com/BurntSushi/toml v1.5.0 // indirect
+	github.com/Masterminds/semver/v3 v3.4.0 // indirect
+	github.com/Microsoft/go-winio v0.6.2 // indirect
 	github.com/beorn7/perks v1.0.1 // indirect
 	github.com/cespare/xxhash/v2 v2.3.0 // indirect
 	github.com/containerd/errdefs v1.0.0 // indirect
 	github.com/containerd/errdefs/pkg v0.3.0 // indirect
 	github.com/containerd/log v0.1.0 // indirect
 	github.com/coreos/go-semver v0.3.1 // indirect
-	github.com/coreos/go-systemd/v22 v22.5.0 // indirect
+	github.com/coreos/go-systemd/v22 v22.7.0 // indirect
+	github.com/davecgh/go-spew v1.1.2-0.20180830191138-d8f796af33cc // indirect
 	github.com/dgryski/go-farm v0.0.0-20240924180020-3414d57e47da // indirect
 	github.com/distribution/reference v0.6.0 // indirect
-	github.com/dlclark/regexp2 v1.11.4 // indirect
-	github.com/docker/go-connections v0.5.0 // indirect
+	github.com/docker/go-connections v0.7.0 // indirect
 	github.com/docker/go-units v0.5.0 // indirect
 	github.com/eapache/channels v1.1.0 // indirect
 	github.com/eapache/queue v1.1.0 // indirect
-	github.com/emicklei/go-restful/v3 v3.12.1 // indirect
+	github.com/emicklei/go-restful/v3 v3.13.0 // indirect
 	github.com/evanphx/json-patch/v5 v5.9.0 // indirect
 	github.com/felixge/httpsnoop v1.0.4 // indirect
-	github.com/fsnotify/fsnotify v1.8.0 // indirect
-	github.com/fxamacker/cbor/v2 v2.7.0 // indirect
+	github.com/fsnotify/fsnotify v1.9.0 // indirect
+	github.com/fxamacker/cbor/v2 v2.9.0 // indirect
+	github.com/gaissmai/bart v0.26.1 // indirect
 	github.com/go-logr/logr v1.4.3 // indirect
 	github.com/go-logr/stdr v1.2.2 // indirect
 	github.com/go-openapi/jsonpointer v0.21.0 // indirect
 	github.com/go-openapi/jsonreference v0.21.0 // indirect
 	github.com/go-openapi/swag v0.23.0 // indirect
 	github.com/go-task/slim-sprig/v3 v3.0.0 // indirect
-	github.com/gogo/protobuf v1.3.2 // indirect
+	github.com/go-viper/mapstructure/v2 v2.4.0 // indirect
 	github.com/golang/protobuf v1.5.4 // indirect
-	github.com/google/gnostic-models v0.6.9 // indirect
-	github.com/google/pprof v0.0.0-20250403155104-27863c87afa6 // indirect
+	github.com/google/gnostic-models v0.7.0 // indirect
+	github.com/google/pprof v0.0.0-20260402051712-545e8a4df936 // indirect
 	github.com/google/uuid v1.6.0 // indirect
-	github.com/gookit/color v1.5.4 // indirect
-	github.com/gookit/goutil v0.6.18 // indirect
-	github.com/gookit/gsr v0.1.0 // indirect
-	github.com/grpc-ecosystem/grpc-gateway/v2 v2.27.1 // indirect
-	github.com/hashicorp/hcl v1.0.0 // indirect
+	github.com/gookit/color v1.6.1 // indirect
+	github.com/gookit/goutil v0.7.6 // indirect
+	github.com/gookit/gsr v0.1.1 // indirect
+	github.com/grpc-ecosystem/grpc-gateway/v2 v2.29.0 // indirect
 	github.com/inconshreveable/mousetrap v1.1.0 // indirect
 	github.com/josharian/intern v1.0.0 // indirect
 	github.com/josharian/native v1.1.0 // indirect
 	github.com/json-iterator/go v1.1.12 // indirect
 	github.com/k-sone/critbitgo v1.4.0 // indirect
-	github.com/magiconair/properties v1.8.9 // indirect
+	github.com/kylelemons/godebug v1.1.0 // indirect
 	github.com/mailru/easyjson v0.9.0 // indirect
 	github.com/mattn/go-isatty v0.0.20 // indirect
 	github.com/mdlayher/genetlink v1.3.2 // indirect
-	github.com/mdlayher/netlink v1.7.3-0.20250113171957-fbb4dce95f42 // indirect
+	github.com/mdlayher/netlink v1.8.0 // indirect
 	github.com/mdlayher/packet v1.1.2 // indirect
 	github.com/mdlayher/socket v0.5.1 // indirect
-	github.com/mitchellh/mapstructure v1.5.0 // indirect
 	github.com/moby/docker-image-spec v1.3.1 // indirect
 	github.com/moby/sys/atomicwriter v0.1.0 // indirect
 	github.com/moby/term v0.5.2 // indirect
 	github.com/modern-go/concurrent v0.0.0-20180306012644-bacd9c7ef1dd // indirect
-	github.com/modern-go/reflect2 v1.0.2 // indirect
-	github.com/morikuni/aec v1.0.0 // indirect
+	github.com/modern-go/reflect2 v1.0.3-0.20250322232337-35a7c28c31ee // indirect
+	github.com/morikuni/aec v1.1.0 // indirect
 	github.com/munnerz/goautoneg v0.0.0-20191010083416-a7dc8b61c822 // indirect
 	github.com/opencontainers/go-digest v1.0.0 // indirect
 	github.com/opencontainers/image-spec v1.1.1 // indirect
+	github.com/orcaman/concurrent-map/v2 v2.0.1 // indirect
 	github.com/pelletier/go-toml v1.9.5 // indirect
 	github.com/pelletier/go-toml/v2 v2.2.3 // indirect
 	github.com/pierrec/lz4/v4 v4.1.22 // indirect
 	github.com/pmezard/go-difflib v1.0.1-0.20181226105442-5d4384ee4fb2 // indirect
-	github.com/prometheus/client_model v0.6.1 // indirect
-	github.com/prometheus/common v0.62.0 // indirect
-	github.com/prometheus/procfs v0.15.1 // indirect
-	github.com/remyoudompheng/bigfft v0.0.0-20230129092748-24d4a6f8daec // indirect
-	github.com/sagikazarmark/locafero v0.6.0 // indirect
-	github.com/sagikazarmark/slog-shim v0.1.0 // indirect
+	github.com/prometheus/client_model v0.6.2 // indirect
+	github.com/prometheus/common v0.70.1 // indirect
+	github.com/prometheus/procfs v0.21.1 // indirect
+	github.com/sagikazarmark/locafero v0.7.0 // indirect
+	github.com/segmentio/fasthash v1.0.3 // indirect
 	github.com/sourcegraph/conc v0.3.0 // indirect
-	github.com/spf13/afero v1.11.0 // indirect
+	github.com/spf13/afero v1.12.0 // indirect
 	github.com/spf13/cast v1.7.1 // indirect
-	github.com/spf13/pflag v1.0.6 // indirect
-	github.com/spf13/viper v1.19.0 // indirect
+	github.com/spf13/pflag v1.0.10 // indirect
+	github.com/spf13/viper v1.20.1 // indirect
 	github.com/subosito/gotenv v1.6.0 // indirect
-	github.com/tj/go-spin v1.1.0 // indirect
 	github.com/u-root/uio v0.0.0-20240224005618-d2acac8f3701 // indirect
 	github.com/valyala/bytebufferpool v1.0.0 // indirect
 	github.com/vishvananda/netns v0.0.5 // indirect
 	github.com/x448/float16 v0.8.4 // indirect
-	github.com/xlab/c-for-go v1.3.0 // indirect
-	github.com/xlab/pkgconfig v0.0.0-20170226114623-cea12a0fd245 // indirect
 	github.com/xo/terminfo v0.0.0-20220910002029-abceb7e1c41e // indirect
-	go.opentelemetry.io/auto/sdk v1.1.0 // indirect
-	go.opentelemetry.io/contrib/instrumentation/net/http/otelhttp v0.49.0 // indirect
-	go.opentelemetry.io/otel v1.37.0 // indirect
-	go.opentelemetry.io/otel/exporters/otlp/otlptrace/otlptracehttp v1.37.0 // indirect
-	go.opentelemetry.io/otel/metric v1.37.0 // indirect
-	go.opentelemetry.io/otel/sdk v1.37.0 // indirect
-	go.opentelemetry.io/otel/trace v1.37.0 // indirect
-	go.uber.org/automaxprocs v1.6.0 // indirect
+	go.opentelemetry.io/auto/sdk v1.2.1 // indirect
+	go.opentelemetry.io/contrib/instrumentation/net/http/otelhttp v0.61.0 // indirect
+	go.opentelemetry.io/otel v1.44.0 // indirect
+	go.opentelemetry.io/otel/exporters/otlp/otlptrace/otlptracehttp v1.43.0 // indirect
+	go.opentelemetry.io/otel/metric v1.44.0 // indirect
+	go.opentelemetry.io/otel/trace v1.44.0 // indirect
 	go.uber.org/multierr v1.11.0 // indirect
-	golang.org/x/crypto v0.39.0 // indirect
-	golang.org/x/mod v0.25.0 // indirect
-	golang.org/x/net v0.41.0 // indirect
-	golang.org/x/oauth2 v0.30.0 // indirect
-	golang.org/x/term v0.32.0 // indirect
-	golang.org/x/text v0.26.0 // indirect
-	golang.org/x/time v0.9.0 // indirect
-	golang.org/x/tools v0.33.0 // indirect
+	go.yaml.in/yaml/v2 v2.4.4 // indirect
+	go.yaml.in/yaml/v3 v3.0.4 // indirect
+	golang.org/x/crypto v0.54.0 // indirect
+	golang.org/x/mod v0.38.0 // indirect
+	golang.org/x/net v0.57.0 // indirect
+	golang.org/x/oauth2 v0.36.0 // indirect
+	golang.org/x/term v0.45.0 // indirect
+	golang.org/x/text v0.40.0 // indirect
+	golang.org/x/time v0.14.0 // indirect
+	golang.org/x/tools v0.48.0 // indirect
 	golang.zx2c4.com/wireguard v0.0.0-20231211153847-12269c276173 // indirect
-	google.golang.org/genproto/googleapis/api v0.0.0-20250603155806-513f23925822 // indirect
-	google.golang.org/genproto/googleapis/rpc v0.0.0-20250603155806-513f23925822 // indirect
-	gopkg.in/evanphx/json-patch.v4 v4.12.0 // indirect
+	google.golang.org/genproto/googleapis/api v0.0.0-20260526163538-3dc84a4a5aaa // indirect
+	google.golang.org/genproto/googleapis/rpc v0.0.0-20260526163538-3dc84a4a5aaa // indirect
+	google.golang.org/protobuf v1.36.12-0.20260120151049-f2248ac996af // indirect
+	gopkg.in/evanphx/json-patch.v4 v4.13.0 // indirect
 	gopkg.in/inf.v0 v0.9.1 // indirect
-	gopkg.in/ini.v1 v1.67.0 // indirect
-	gopkg.in/yaml.v2 v2.4.0 // indirect
-	gopkg.in/yaml.v3 v3.0.1 // indirect
-	k8s.io/kube-openapi v0.0.0-20250318190949-c8a335a9a2ff // indirect
-	k8s.io/utils v0.0.0-20241210054802-24370beab758 // indirect
-	modernc.org/cc/v4 v4.24.4 // indirect
-	modernc.org/mathutil v1.7.1 // indirect
-	modernc.org/opt v0.1.4 // indirect
-	modernc.org/sortutil v1.2.1 // indirect
-	modernc.org/strutil v1.2.1 // indirect
-	modernc.org/token v1.1.0 // indirect
-	sigs.k8s.io/json v0.0.0-20241014173422-cfa47c3a1cc8 // indirect
+	k8s.io/kube-openapi v0.0.0-20260317180543-43fb72c5454a // indirect
+	k8s.io/utils v0.0.0-20260210185600-b8788abfbbc2 // indirect
+	sigs.k8s.io/json v0.0.0-20250730193827-2d320260d730 // indirect
 	sigs.k8s.io/randfill v1.0.0 // indirect
-	sigs.k8s.io/structured-merge-diff/v4 v4.6.0 // indirect
+	sigs.k8s.io/structured-merge-diff/v6 v6.3.3 // indirect
 )

go.sum

@@ -2,32 +2,34 @@ al.essio.dev/pkg/shellescape v1.5.1 h1:86HrALUujYS/h+GtqoB26SBEdkWfmMI6FubjXlsXy
 al.essio.dev/pkg/shellescape v1.5.1/go.mod h1:6sIqp7X2P6mThCQ7twERpZTuigpr6KbZWtls1U8I890=
 github.com/Azure/go-ansiterm v0.0.0-20250102033503-faa5f7b0171c h1:udKWzYgxTojEKWjV8V+WSxDXJ4NFATAsZjh8iIbsQIg=
 github.com/Azure/go-ansiterm v0.0.0-20250102033503-faa5f7b0171c/go.mod h1:xomTg63KZ2rFqZQzSB4Vz2SUXa1BpHTVz9L5PTmPC4E=
-github.com/BurntSushi/toml v0.3.1/go.mod h1:xHWCNGjB5oqiDr8zfno3MHue2Ht5sIBksp03qcyfWMU=
-github.com/BurntSushi/toml v0.4.1/go.mod h1:CxXYINrC8qIiEnFrOxCa7Jy5BFHlXnUU2pbicEuybxQ=
-github.com/BurntSushi/toml v1.4.0 h1:kuoIxZQy2WRRk1pttg9asf+WVv6tWQuBNVmK8+nqPr0=
-github.com/BurntSushi/toml v1.4.0/go.mod h1:ukJfTF/6rtPPRCnwkur4qwRxa8vTRFBF0uk2lLoLwho=
-github.com/Microsoft/go-winio v0.4.14 h1:+hMXMk01us9KgxGb7ftKQt2Xpf5hH/yky+TDA+qxleU=
-github.com/Microsoft/go-winio v0.4.14/go.mod h1:qXqCSQ3Xa7+6tgxaGTIe4Kpcdsi+P8jBhyzoq1bpyYA=
+github.com/BurntSushi/toml v1.5.0 h1:W5quZX/G/csjUnuI8SUYlsHs9M38FC7znL0lIO+DvMg=
+github.com/BurntSushi/toml v1.5.0/go.mod h1:ukJfTF/6rtPPRCnwkur4qwRxa8vTRFBF0uk2lLoLwho=
+github.com/Masterminds/semver/v3 v3.4.0 h1:Zog+i5UMtVoCU8oKka5P7i9q9HgrJeGzI9SA1Xbatp0=
+github.com/Masterminds/semver/v3 v3.4.0/go.mod h1:4V+yj/TJE1HU9XfppCwVMZq3I84lprf4nC11bSS5beM=
+github.com/Microsoft/go-winio v0.6.2 h1:F2VQgta7ecxGYO8k3ZZz3RS8fVIXVxONVUPlNERoyfY=
+github.com/Microsoft/go-winio v0.6.2/go.mod h1:yd8OoFMLzJbo9gZq8j5qaps8bJ9aShtEA8Ipt1oGCvU=
 github.com/beorn7/perks v1.0.1 h1:VlbKKnNfV8bJzeqoa4cOKqO6bYr3WgKZxO8Z16+hsOM=
 github.com/beorn7/perks v1.0.1/go.mod h1:G2ZrVWU2WbWT9wwq4/hrbKbnv/1ERSJQ0ibhJ6rlkpw=
-github.com/cenkalti/backoff/v5 v5.0.2 h1:rIfFVxEf1QsI7E1ZHfp/B4DF/6QBAUhmgkxc0H7Zss8=
-github.com/cenkalti/backoff/v5 v5.0.2/go.mod h1:rkhZdG3JZukswDf7f0cwqPNk4K0sa+F97BxZthm/crw=
+github.com/cenkalti/backoff/v5 v5.0.3 h1:ZN+IMa753KfX5hd8vVaMixjnqRZ3y8CuJKRKj1xcsSM=
+github.com/cenkalti/backoff/v5 v5.0.3/go.mod h1:rkhZdG3JZukswDf7f0cwqPNk4K0sa+F97BxZthm/crw=
 github.com/cespare/xxhash/v2 v2.3.0 h1:UL815xU9SqsFlibzuggzjXhog7bL6oX9BbNZnL2UFvs=
 github.com/cespare/xxhash/v2 v2.3.0/go.mod h1:VGX0DQ3Q6kWi7AoAeZDth3/j3BFtOZR5XLFGgcrjCOs=
-github.com/cilium/ebpf v0.5.0/go.mod h1:4tRaxcgiL706VnOzHOdBlY8IEAIdxINsQBcU4xJJXRs=
-github.com/cilium/ebpf v0.7.0/go.mod h1:/oI2+1shJiTGAMgl6/RgJr36Eo1jzrRcAWbcXO2usCA=
-github.com/cloudflare/ipvs v0.11.0 h1:niLcbqfv8+RSYk+yI+jhiCIHobEdEhO++mcwCX0Znsw=
-github.com/cloudflare/ipvs v0.11.0/go.mod h1:XsbuKcQpqb3rpjYPtsFsM8BNZlgLw+Z0iC6RhocqDd8=
+github.com/cloudflare/ipvs v0.12.0 h1:UeKRM4q82F+XQjw1sGRwsH3IOy6keBYUwlqTUqJhy7Y=
+github.com/cloudflare/ipvs v0.12.0/go.mod h1:SvzXyDX2E33X/Cokbgts7rFXueiSqp6NlEMGDT7Bu24=
 github.com/containerd/errdefs v1.0.0 h1:tg5yIfIlQIrxYtu9ajqY42W3lpS19XqdxRQeEwYG8PI=
 github.com/containerd/errdefs v1.0.0/go.mod h1:+YBYIdtsnF4Iw6nWZhJcqGSg/dwvV7tyJ/kCkyJ2k+M=
 github.com/containerd/errdefs/pkg v0.3.0 h1:9IKJ06FvyNlexW690DXuQNx2KA2cUJXx151Xdx3ZPPE=
 github.com/containerd/errdefs/pkg v0.3.0/go.mod h1:NJw6s9HwNuRhnjJhM7pylWwMyAkmCQvQ4GpJHEqRLVk=
 github.com/containerd/log v0.1.0 h1:TCJt7ioM2cr/tfR8GPbGf9/VRAX8D2B4PjzCpfX540I=
 github.com/containerd/log v0.1.0/go.mod h1:VRRf09a7mHDIRezVKTRCrOq78v577GXq3bSa3EhrzVo=
+github.com/containernetworking/cni v1.3.0 h1:v6EpN8RznAZj9765HhXQrtXgX+ECGebEYEmnuFjskwo=
+github.com/containernetworking/cni v1.3.0/go.mod h1:Bs8glZjjFfGPHMw6hQu82RUgEPNGEaBb9KS5KtNMnJ4=
+github.com/containernetworking/plugins v1.9.1 h1:8oU6WsIsU3bpnNZuvHp74a6cE1MJwbj2P7s4/yTUNlA=
+github.com/containernetworking/plugins v1.9.1/go.mod h1:fj7kS55qg3o/RgS+WGsF3+ZxwIImMPusQZKzBpcSr4c=
 github.com/coreos/go-semver v0.3.1 h1:yi21YpKnrx1gt5R+la8n5WgS0kCrsPp33dmEyHReZr4=
 github.com/coreos/go-semver v0.3.1/go.mod h1:irMmmIw/7yzSRPWryHsK7EYSg09caPQL03VsM8rvUec=
-github.com/coreos/go-systemd/v22 v22.5.0 h1:RrqgGjYQKalulkV8NGVIfkXQf6YYmOyiJKk8iXXhfZs=
-github.com/coreos/go-systemd/v22 v22.5.0/go.mod h1:Y58oyj3AT4RCenI/lSvhwexgC+NSVTIJ3seZv2GcEnc=
+github.com/coreos/go-systemd/v22 v22.7.0 h1:LAEzFkke61DFROc7zNLX/WA2i5J8gYqe0rSj9KI28KA=
+github.com/coreos/go-systemd/v22 v22.7.0/go.mod h1:xNUYtjHu2EDXbsxz1i41wouACIwT7Ybq9o0BQhMwD0w=
 github.com/cpuguy83/go-md2man/v2 v2.0.6/go.mod h1:oOW0eioCTA6cOiMLiUPZOpcVxMig6NIQQ7OS05n1F4g=
 github.com/davecgh/go-spew v1.1.0/go.mod h1:J7Y8YcW2NihsgmVo/mv3lAwl/skON4iLHjSsI+c5H38=
 github.com/davecgh/go-spew v1.1.1/go.mod h1:J7Y8YcW2NihsgmVo/mv3lAwl/skON4iLHjSsI+c5H38=
@@ -37,35 +39,38 @@ github.com/dgryski/go-farm v0.0.0-20240924180020-3414d57e47da h1:aIftn67I1fkbMa5
 github.com/dgryski/go-farm v0.0.0-20240924180020-3414d57e47da/go.mod h1:SqUrOPUnsFjfmXRMNPybcSiG0BgUW2AuFH8PAnS2iTw=
 github.com/distribution/reference v0.6.0 h1:0IXCQ5g4/QMHHkarYzh5l+u8T3t73zM5QvfrDyIgxBk=
 github.com/distribution/reference v0.6.0/go.mod h1:BbU0aIcezP1/5jX/8MP0YiH4SdvB5Y4f/wlDRiLyi3E=
-github.com/dlclark/regexp2 v1.11.4 h1:rPYF9/LECdNymJufQKmri9gV604RvvABwgOA8un7yAo=
-github.com/dlclark/regexp2 v1.11.4/go.mod h1:DHkYz0B9wPfa6wondMfaivmHpzrQ3v9q8cnmRbL6yW8=
-github.com/docker/docker v28.3.3+incompatible h1:Dypm25kh4rmk49v1eiVbsAtpAsYURjYkaKubwuBdxEI=
-github.com/docker/docker v28.3.3+incompatible/go.mod h1:eEKB0N0r5NX/I1kEveEz05bcu8tLC/8azJZsviup8Sk=
-github.com/docker/go-connections v0.5.0 h1:USnMq7hx7gwdVZq1L49hLXaFtUdTADjXGp+uj1Br63c=
-github.com/docker/go-connections v0.5.0/go.mod h1:ov60Kzw0kKElRwhNs9UlUHAE/F9Fe6GLaXnqyDdmEXc=
+github.com/docker/docker v28.5.2+incompatible h1:DBX0Y0zAjZbSrm1uzOkdr1onVghKaftjlSWt4AFexzM=
+github.com/docker/docker v28.5.2+incompatible/go.mod h1:eEKB0N0r5NX/I1kEveEz05bcu8tLC/8azJZsviup8Sk=
+github.com/docker/go-connections v0.7.0 h1:6SsRfJddP22WMrCkj19x9WKjEDTB+ahsdiGYf0mN39c=
+github.com/docker/go-connections v0.7.0/go.mod h1:no1qkHdjq7kLMGUXYAduOhYPSJxxvgWBh7ogVvptn3Q=
 github.com/docker/go-units v0.5.0 h1:69rxXcBk27SvSaaxTtLh/8llcHD8vYHT7WSdRZ/jvr4=
 github.com/docker/go-units v0.5.0/go.mod h1:fgPhTUdO+D/Jk86RDLlptpiXQzgHJF7gydDDbaIK4Dk=
-github.com/dustin/go-humanize v1.0.1 h1:GzkhY7T5VNhEkwH0PVJgjz+fX1rhBrR7pRT3mDkpeCY=
-github.com/dustin/go-humanize v1.0.1/go.mod h1:Mu1zIs6XwVuF/gI1OepvI0qD18qycQx+mFykh5fBlto=
 github.com/eapache/channels v1.1.0 h1:F1taHcn7/F0i8DYqKXJnyhJcVpp2kgFcNePxXtnyu4k=
 github.com/eapache/channels v1.1.0/go.mod h1:jMm2qB5Ubtg9zLd+inMZd2/NUvXgzmWXsDaLyQIGfH0=
 github.com/eapache/queue v1.1.0 h1:YOEu7KNc61ntiQlcEeUIoDTJ2o8mQznoNvUhiigpIqc=
 github.com/eapache/queue v1.1.0/go.mod h1:6eCeP0CKFpHLu8blIFXhExK/dRa7WDZfr6jVFPTqq+I=
-github.com/emicklei/go-restful/v3 v3.12.1 h1:PJMDIM/ak7btuL8Ex0iYET9hxM3CI2sjZtzpL63nKAU=
-github.com/emicklei/go-restful/v3 v3.12.1/go.mod h1:6n3XBCmQQb25CM2LCACGz8ukIrRry+4bhvbpWn3mrbc=
+github.com/emicklei/go-restful/v3 v3.13.0 h1:C4Bl2xDndpU6nJ4bc1jXd+uTmYPVUwkD6bFY/oTyCes=
+github.com/emicklei/go-restful/v3 v3.13.0/go.mod h1:6n3XBCmQQb25CM2LCACGz8ukIrRry+4bhvbpWn3mrbc=
 github.com/evanphx/json-patch/v5 v5.9.0 h1:kcBlZQbplgElYIlo/n1hJbls2z/1awpXxpRi0/FOJfg=
 github.com/evanphx/json-patch/v5 v5.9.0/go.mod h1:VNkHZ/282BpEyt/tObQO8s5CMPmYYq14uClGH4abBuQ=
 github.com/felixge/httpsnoop v1.0.4 h1:NFTV2Zj1bL4mc9sqWACXbQFVBBg2W3GPvqp8/ESS2Wg=
 github.com/felixge/httpsnoop v1.0.4/go.mod h1:m8KPJKqk1gH5J9DgRY2ASl2lWCfGKXixSwevea8zH2U=
-github.com/florianl/go-conntrack v0.4.0 h1:TlYkxytdwgVayfU0cKwkHurQA0Rd1ZSEBRckRYDUu18=
-github.com/florianl/go-conntrack v0.4.0/go.mod h1:iPDx4oIats2T7X7Jm3PFyRCJM1GfZhJaSHOWROYOrE8=
-github.com/frankban/quicktest v1.11.3/go.mod h1:wRf/ReqHper53s+kmmSZizM8NamnL3IM0I9ntUbOk+k=
+github.com/florianl/go-conntrack v0.7.0 h1:kWbRhLRUqBlmi2ncASmygN5yI281wlpnTUJs69tcs7g=
+github.com/florianl/go-conntrack v0.7.0/go.mod h1:7kzfTjtQCRsAVkuornpI8KiwFfa79QmcZXZQex1IXhU=
 github.com/frankban/quicktest v1.14.6 h1:7Xjx+VpznH+oBnejlPUj8oUpdxnVs4f8XU8WnHkI4W8=
 github.com/frankban/quicktest v1.14.6/go.mod h1:4ptaffx2x8+WTWXmUCuVU6aPUX1/Mz7zb5vbUoiM6w0=
-github.com/fsnotify/fsnotify v1.8.0 h1:dAwr6QBTBZIkG8roQaJjGof0pp0EeF+tNV7YBP3F/8M=
-github.com/fsnotify/fsnotify v1.8.0/go.mod h1:8jBTzvmWwFyi3Pb8djgCCO5IBqzKJ/Jwo8TRcHyHii0=
-github.com/fxamacker/cbor/v2 v2.7.0 h1:iM5WgngdRBanHcxugY4JySA0nk1wZorNOpTgCMedv5E=
-github.com/fxamacker/cbor/v2 v2.7.0/go.mod h1:pxXPTn3joSm21Gbwsv0w9OSA2y1HFR9qXEeXQVeNoDQ=
+github.com/fsnotify/fsnotify v1.9.0 h1:2Ml+OJNzbYCTzsxtv8vKSFD9PbJjmhYF14k/jKC7S9k=
+github.com/fsnotify/fsnotify v1.9.0/go.mod h1:8jBTzvmWwFyi3Pb8djgCCO5IBqzKJ/Jwo8TRcHyHii0=
+github.com/fxamacker/cbor/v2 v2.9.0 h1:NpKPmjDBgUfBms6tr6JZkTHtfFGcMKsw3eGcmD/sapM=
+github.com/fxamacker/cbor/v2 v2.9.0/go.mod h1:vM4b+DJCtHn+zz7h3FFp/hDAI9WNWCsZj23V5ytsSxQ=
+github.com/gaissmai/bart v0.26.1 h1:+w4rnLGNlA2GDVn382Tfe3jOsK5vOr5n4KmigJ9lbTo=
+github.com/gaissmai/bart v0.26.1/go.mod h1:GREWQfTLRWz/c5FTOsIw+KkscuFkIV5t8Rp7Nd1Td5c=
+github.com/gkampitakis/ciinfo v0.3.2 h1:JcuOPk8ZU7nZQjdUhctuhQofk7BGHuIy0c9Ez8BNhXs=
+github.com/gkampitakis/ciinfo v0.3.2/go.mod h1:1NIwaOcFChN4fa/B0hEBdAb6npDlFL8Bwx4dfRLRqAo=
+github.com/gkampitakis/go-diff v1.3.2 h1:Qyn0J9XJSDTgnsgHRdz9Zp24RaJeKMUHg2+PDZZdC4M=
+github.com/gkampitakis/go-diff v1.3.2/go.mod h1:LLgOrpqleQe26cte8s36HTWcTmMEur6OPYerdAAS9tk=
+github.com/gkampitakis/go-snaps v0.5.15 h1:amyJrvM1D33cPHwVrjo9jQxX8g/7E2wYdZ+01KS3zGE=
+github.com/gkampitakis/go-snaps v0.5.15/go.mod h1:HNpx/9GoKisdhw9AFOBT1N7DBs9DiHo/hGheFGBZ+mc=
 github.com/go-logr/logr v1.2.2/go.mod h1:jdQByPbusPIv2/zmleS9BjJVeZ6kBagPoEUsqbVz/1A=
 github.com/go-logr/logr v1.4.3 h1:CjnDlHq8ikf6E492q6eKboGOC0T8CDaOvkHCIg8idEI=
 github.com/go-logr/logr v1.4.3/go.mod h1:9T104GzyrTigFIr8wt5mBrctHMim0Nb2HLGrmQ40KvY=
@@ -79,125 +84,96 @@ github.com/go-openapi/swag v0.23.0 h1:vsEVJDUo2hPJ2tu0/Xc+4noaxyEffXNIs3cOULZ+Gr
 github.com/go-openapi/swag v0.23.0/go.mod h1:esZ8ITTYEsH1V2trKHjAN8Ai7xHb8RV+YSZ577vPjgQ=
 github.com/go-task/slim-sprig/v3 v3.0.0 h1:sUs3vkvUymDpBKi3qH1YSqBQk9+9D/8M2mN1vB6EwHI=
 github.com/go-task/slim-sprig/v3 v3.0.0/go.mod h1:W848ghGpv3Qj3dhTPRyJypKRiqCdHZiAzKg9hl15HA8=
-github.com/go-test/deep v1.1.0 h1:WOcxcdHcvdgThNXjw0t76K42FXTU7HpNQWHpA2HHNlg=
-github.com/go-test/deep v1.1.0/go.mod h1:5C2ZWiW0ErCdrYzpqxLbTX7MG14M9iiw8DgHncVwcsE=
-github.com/godbus/dbus/v5 v5.0.4/go.mod h1:xhWf0FNVPg57R7Z0UbKHbJfkEywrmjJnf7w5xrFpKfA=
-github.com/gogo/protobuf v1.3.2 h1:Ov1cvc58UF3b5XjBnZv7+opcTcQFZebYjWzi34vdm4Q=
-github.com/gogo/protobuf v1.3.2/go.mod h1:P1XiOD3dCwIKUDQYPy72D8LYyHL2YPYrpS2s69NZV8Q=
+github.com/go-test/deep v1.1.1 h1:0r/53hagsehfO4bzD2Pgr/+RgHqhmf+k1Bpse2cTu1U=
+github.com/go-test/deep v1.1.1/go.mod h1:5C2ZWiW0ErCdrYzpqxLbTX7MG14M9iiw8DgHncVwcsE=
+github.com/go-viper/mapstructure/v2 v2.4.0 h1:EBsztssimR/CONLSZZ04E8qAkxNYq4Qp9LvH92wZUgs=
+github.com/go-viper/mapstructure/v2 v2.4.0/go.mod h1:oJDH3BJKyqBA2TXFhDsKDGDTlndYOZ6rGS0BRZIxGhM=
+github.com/goccy/go-yaml v1.18.0 h1:8W7wMFS12Pcas7KU+VVkaiCng+kG8QiFeFwzFb+rwuw=
+github.com/goccy/go-yaml v1.18.0/go.mod h1:XBurs7gK8ATbW4ZPGKgcbrY1Br56PdM69F7LkFRi1kA=
 github.com/golang/protobuf v1.5.4 h1:i7eJL8qZTpSEXOPTxNKhASYpMn+8e5Q6AdndVa1dWek=
 github.com/golang/protobuf v1.5.4/go.mod h1:lnTiLA8Wa4RWRcIUkrtSVa5nRhsEGBg48fD6rSs7xps=
-github.com/google/gnostic-models v0.6.9 h1:MU/8wDLif2qCXZmzncUQ/BOfxWfthHi63KqpoNbWqVw=
-github.com/google/gnostic-models v0.6.9/go.mod h1:CiWsm0s6BSQd1hRn8/QmxqB6BesYcbSZxsz9b0KuDBw=
-github.com/google/go-cmp v0.2.0/go.mod h1:oXzfMopK8JAjlY9xF4vHSVASa0yLyX7SntLO5aqRK0M=
-github.com/google/go-cmp v0.3.1/go.mod h1:8QqcDgzrUqlUb/G2PQTWiueGozuR1884gddMywk6iLU=
-github.com/google/go-cmp v0.4.0/go.mod h1:v8dTdLbMG2kIc/vJvl+f65V22dbkXbowE6jgT/gNBxE=
-github.com/google/go-cmp v0.5.2/go.mod h1:v8dTdLbMG2kIc/vJvl+f65V22dbkXbowE6jgT/gNBxE=
-github.com/google/go-cmp v0.5.4/go.mod h1:v8dTdLbMG2kIc/vJvl+f65V22dbkXbowE6jgT/gNBxE=
-github.com/google/go-cmp v0.5.5/go.mod h1:v8dTdLbMG2kIc/vJvl+f65V22dbkXbowE6jgT/gNBxE=
-github.com/google/go-cmp v0.5.6/go.mod h1:v8dTdLbMG2kIc/vJvl+f65V22dbkXbowE6jgT/gNBxE=
+github.com/google/gnostic-models v0.7.0 h1:qwTtogB15McXDaNqTZdzPJRHvaVJlAl+HVQnLmJEJxo=
+github.com/google/gnostic-models v0.7.0/go.mod h1:whL5G0m6dmc5cPxKc5bdKdEN3UjI7OUGxBlw57miDrQ=
+github.com/google/go-cmp v0.5.7/go.mod h1:n+brtR0CgQNWTVd5ZUFpTBC8YFBDLK/h/bpaJ8/DtOE=
 github.com/google/go-cmp v0.5.9/go.mod h1:17dUlkBOakJ0+DkrSSNjCkIjxS6bF9zb3elmeNGIjoY=
 github.com/google/go-cmp v0.7.0 h1:wk8382ETsv4JYUZwIsn6YpYiWiBsYLSJiTsyBybVuN8=
 github.com/google/go-cmp v0.7.0/go.mod h1:pXiqmnSA92OHEEa9HXL2W4E7lf9JzCmGVUdgjX3N/iU=
+github.com/google/go-containerregistry v0.21.9 h1:F+D4uZ3iA3DLMJLfhaqMdHJbzeqm/216WGQq2dokuLs=
+github.com/google/go-containerregistry v0.21.9/go.mod h1:dP5XNKcL7kMFF/TB3LfvWmVhAcv7iqkHb3oDK8aauTo=
 github.com/google/gofuzz v1.0.0/go.mod h1:dBl0BpW6vV/+mYPU4Po3pmUjxk6FQPldtuIdl/M65Eg=
 github.com/google/nftables v0.3.0 h1:bkyZ0cbpVeMHXOrtlFc8ISmfVqq5gPJukoYieyVmITg=
 github.com/google/nftables v0.3.0/go.mod h1:BCp9FsrbF1Fn/Yu6CLUc9GGZFw/+hsxfluNXXmxBfRM=
-github.com/google/pprof v0.0.0-20250403155104-27863c87afa6 h1:BHT72Gu3keYf3ZEu2J0b1vyeLSOYI8bm5wbJM/8yDe8=
-github.com/google/pprof v0.0.0-20250403155104-27863c87afa6/go.mod h1:boTsfXsheKC2y+lKOCMpSfarhxDeIzfZG1jqGcPl3cA=
+github.com/google/pprof v0.0.0-20260402051712-545e8a4df936 h1:EwtI+Al+DeppwYX2oXJCETMO23COyaKGP6fHVpkpWpg=
+github.com/google/pprof v0.0.0-20260402051712-545e8a4df936/go.mod h1:MxpfABSjhmINe3F1It9d+8exIHFvUqtLIRCdOGNXqiI=
 github.com/google/shlex v0.0.0-20191202100458-e7afc7fbc510 h1:El6M4kTTCOh6aBiKaUGG7oYTSPP8MxqL4YI3kZKwcP4=
 github.com/google/shlex v0.0.0-20191202100458-e7afc7fbc510/go.mod h1:pupxD2MaaD3pAXIBCelhxNneeOaAeabZDe5s4K6zSpQ=
 github.com/google/uuid v1.6.0 h1:NIvaJDMOsjHA8n1jAhLSgzrAzy1Hgr+hNrb57e+94F0=
 github.com/google/uuid v1.6.0/go.mod h1:TIyPZe4MgqvfeYDBFedMoGGpEw/LqOeaOT+nhxU+yHo=
-github.com/gookit/color v1.5.4 h1:FZmqs7XOyGgCAxmWyPslpiok1k05wmY3SJTytgvYFs0=
-github.com/gookit/color v1.5.4/go.mod h1:pZJOeOS8DM43rXbp4AZo1n9zCU2qjpcRko0b6/QJi9w=
-github.com/gookit/goutil v0.6.18 h1:MUVj0G16flubWT8zYVicIuisUiHdgirPAkmnfD2kKgw=
-github.com/gookit/goutil v0.6.18/go.mod h1:AY/5sAwKe7Xck+mEbuxj0n/bc3qwrGNe3Oeulln7zBA=
-github.com/gookit/gsr v0.1.0 h1:0gadWaYGU4phMs0bma38t+Do5OZowRMEVlHv31p0Zig=
-github.com/gookit/gsr v0.1.0/go.mod h1:7wv4Y4WCnil8+DlDYHBjidzrEzfHhXEoFjEA0pPPWpI=
-github.com/gookit/slog v0.5.8 h1:XZCeHLQvvOZWcSUDZcqxXITsL9+d1ESsKZoASBmK1lI=
-github.com/gookit/slog v0.5.8/go.mod h1:s0ViFOY/IgUuT4MDPF0l9x5/npcciy8pL4xwWZadnoc=
-github.com/grpc-ecosystem/grpc-gateway/v2 v2.27.1 h1:X5VWvz21y3gzm9Nw/kaUeku/1+uBhcekkmy4IkffJww=
-github.com/grpc-ecosystem/grpc-gateway/v2 v2.27.1/go.mod h1:Zanoh4+gvIgluNqcfMVTJueD4wSS5hT7zTt4Mrutd90=
-github.com/hashicorp/hcl v1.0.0 h1:0Anlzjpi4vEasTeNFn2mLJgTSwt0+6sfsiTG8qcWGx4=
-github.com/hashicorp/hcl v1.0.0/go.mod h1:E5yfLk+7swimpb2L/Alb/PJmXilQ/rhwaUYs4T20WEQ=
+github.com/gookit/assert v0.1.1 h1:lh3GcawXe/p+cU7ESTZ5Ui3Sm/x8JWpIis4/1aF0mY0=
+github.com/gookit/assert v0.1.1/go.mod h1:jS5bmIVQZTIwk42uXl4lyj4iaaxx32tqH16CFj0VX2E=
+github.com/gookit/color v1.6.1 h1:KoTnDxJPRgrL0SoX0f8rCFg2zI0t4E3GZZBMo2nN8LU=
+github.com/gookit/color v1.6.1/go.mod h1:9ACFc7/1IpHGBW8RwuDm/0YEnhg3dwwXpoMsmtyHfjs=
+github.com/gookit/goutil v0.7.6 h1:700ZP6QPWhw5ms7X13JH9fUs4LTyYMmncFFMGpK73ns=
+github.com/gookit/goutil v0.7.6/go.mod h1:vJS9HXctYTCLtCsZot5L5xF+O1oR17cDYO9R0HxBmnU=
+github.com/gookit/gsr v0.1.1 h1:TaHD3M7qa6lcAf9D2J4mGNg+QjgDtD1bw7uctF8RXOM=
+github.com/gookit/gsr v0.1.1/go.mod h1:7wv4Y4WCnil8+DlDYHBjidzrEzfHhXEoFjEA0pPPWpI=
+github.com/gookit/rotatefile v0.3.0 h1:9MtCRBM79/Chcqp6ySHHmeDGpJE09WvyRFHo7yCQ+is=
+github.com/gookit/rotatefile v0.3.0/go.mod h1:MUaLyw2tEKNe8nta7o2qMfCGST30kzJqybG4KUreIu4=
+github.com/gookit/slog v0.7.1 h1:/q4YtsaJfdtzK+Q1g3QtNoO8cuEF6EjjU56W5U069i8=
+github.com/gookit/slog v0.7.1/go.mod h1:aJ4SGHlMR5YdfeQcICQBEn5bnF0Rpnuh+a5FEzQqXpE=
+github.com/grpc-ecosystem/grpc-gateway/v2 v2.29.0 h1:5VipnvEpbqr2gA2VbM+nYVbkIF28c5ZQfqCBQ5g2xfk=
+github.com/grpc-ecosystem/grpc-gateway/v2 v2.29.0/go.mod h1:Hyl3n6Twe1hvtd9XUXDec4pTvgMSEixRuQKPTMH2bNs=
 github.com/hugelgupf/socketpair v0.0.0-20190730060125-05d35a94e714 h1:/jC7qQFrv8CrSJVmaolDVOxTfS9kc36uB6H40kdbQq8=
 github.com/hugelgupf/socketpair v0.0.0-20190730060125-05d35a94e714/go.mod h1:2Goc3h8EklBH5mspfHFxBnEoURQCGzQQH1ga9Myjvis=
 github.com/huin/goupnp v1.3.0 h1:UvLUlWDNpoUdYzb2TCn+MuTWtcjXKSza2n6CBdQ0xXc=
 github.com/huin/goupnp v1.3.0/go.mod h1:gnGPsThkYa7bFi/KWmEysQRf48l2dvR5bxr2OFckNX8=
 github.com/inconshreveable/mousetrap v1.1.0 h1:wN+x4NVGpMsO7ErUn/mUI3vEoE6Jt13X2s0bqwp9tc8=
 github.com/inconshreveable/mousetrap v1.1.0/go.mod h1:vpF70FUmC8bwa3OWnCshd2FqLfsEA9PFc4w1p2J65bw=
-github.com/insomniacslk/dhcp v0.0.0-20241224095048-b56fa0d5f25d h1:VkCNWh6tuQLgDBc6KrUOz/L1mCUQGnR1Ujj8uTgpwwk=
-github.com/insomniacslk/dhcp v0.0.0-20241224095048-b56fa0d5f25d/go.mod h1:VvGYjkZoJyKqlmT1yzakUs4mfKMNB0XdODP0+rdml6k=
+github.com/insomniacslk/dhcp v0.0.0-20260719225207-c76316d4aa82 h1:y5aU8Uvl7eyM5WNgdQvRxbMJb+zo7pD+S72/Yo4pvnQ=
+github.com/insomniacslk/dhcp v0.0.0-20260719225207-c76316d4aa82/go.mod h1:qfvBmyDNp+/liLEYWRvqny/PEz9hGe2Dz833eXILSmo=
 github.com/josharian/intern v1.0.0 h1:vlS4z54oSdjm0bgjRigI+G1HpF+tI+9rE5LLzOg8HmY=
 github.com/josharian/intern v1.0.0/go.mod h1:5DoeVV0s6jJacbCEi61lwdGj/aVlrQvzHFFd8Hwg//Y=
-github.com/josharian/native v0.0.0-20200817173448-b6b71def0850/go.mod h1:7X/raswPFr05uY3HiLlYeyQntB6OO7E/d2Cu7qoaN2w=
+github.com/josharian/native v1.0.0/go.mod h1:7X/raswPFr05uY3HiLlYeyQntB6OO7E/d2Cu7qoaN2w=
 github.com/josharian/native v1.1.0 h1:uuaP0hAbW7Y4l0ZRQ6C9zfb7Mg1mbFKry/xzDAfmtLA=
 github.com/josharian/native v1.1.0/go.mod h1:7X/raswPFr05uY3HiLlYeyQntB6OO7E/d2Cu7qoaN2w=
+github.com/joshdk/go-junit v1.0.0 h1:S86cUKIdwBHWwA6xCmFlf3RTLfVXYQfvanM5Uh+K6GE=
+github.com/joshdk/go-junit v1.0.0/go.mod h1:TiiV0PqkaNfFXjEiyjWM3XXrhVyCa1K4Zfga6W52ung=
 github.com/jpillora/backoff v1.0.0 h1:uvFg412JmmHBHw7iwprIxkPMI+sGQ4kzOWsMeHnm2EA=
 github.com/jpillora/backoff v1.0.0/go.mod h1:J/6gKK9jxlEcS3zixgDgUAsiuZ7yrSoa/FX5e0EB2j4=
-github.com/jsimonetti/rtnetlink v0.0.0-20190606172950-9527aa82566a/go.mod h1:Oz+70psSo5OFh8DBl0Zv2ACw7Esh6pPUphlvZG9x7uw=
-github.com/jsimonetti/rtnetlink v0.0.0-20200117123717-f846d4f6c1f4/go.mod h1:WGuG/smIU4J/54PblvSbh+xvCZmpJnFgr3ds6Z55XMQ=
-github.com/jsimonetti/rtnetlink v0.0.0-20201009170750-9c6f07d100c1/go.mod h1:hqoO/u39cqLeBLebZ8fWdE96O7FxrAsRYhnVOdgHxok=
-github.com/jsimonetti/rtnetlink v0.0.0-20201216134343-bde56ed16391/go.mod h1:cR77jAZG3Y3bsb8hF6fHJbFoyFukLFOkQ98S0pQz3xw=
-github.com/jsimonetti/rtnetlink v0.0.0-20201220180245-69540ac93943/go.mod h1:z4c53zj6Eex712ROyh8WI0ihysb5j2ROyV42iNogmAs=
-github.com/jsimonetti/rtnetlink v0.0.0-20210122163228-8d122574c736/go.mod h1:ZXpIyOK59ZnN7J0BV99cZUPmsqDRZ3eq5X+st7u/oSA=
-github.com/jsimonetti/rtnetlink v0.0.0-20210212075122-66c871082f2b/go.mod h1:8w9Rh8m+aHZIG69YPGGem1i5VzoyRC8nw2kA8B+ik5U=
-github.com/jsimonetti/rtnetlink v0.0.0-20210525051524-4cc836578190/go.mod h1:NmKSdU4VGSiv1bMsdqNALI4RSvvjtz65tTMCnD05qLo=
-github.com/jsimonetti/rtnetlink v0.0.0-20211022192332-93da33804786/go.mod h1:v4hqbTdfQngbVSZJVWUhGE/lbTFf9jb+ygmNUDQMuOs=
 github.com/json-iterator/go v1.1.12 h1:PV8peI4a0ysnczrg+LtxykD8LfKY9ML6u2jnxaEnrnM=
 github.com/json-iterator/go v1.1.12/go.mod h1:e30LSqwooZae/UwlEbR2852Gd8hjQvJoHmT4TnhNGBo=
 github.com/k-sone/critbitgo v1.4.0 h1:l71cTyBGeh6X5ATh6Fibgw3+rtNT80BA0uNNWgkPrbE=
 github.com/k-sone/critbitgo v1.4.0/go.mod h1:7E6pyoyADnFxlUBEKcnfS49b7SUAQGMK+OAp/UQvo0s=
-github.com/kisielk/errcheck v1.5.0/go.mod h1:pFxgyoBC7bSaBwPgfKdkLd5X25qrDl4LWUI2bnpBCr8=
-github.com/kisielk/gotool v1.0.0/go.mod h1:XhKaO+MFFWcvkIS/tQcRk01m1F5IRFswLeQ+oQHNcck=
-github.com/klauspost/compress v1.18.0 h1:c/Cqfb0r+Yi+JtIEq73FWXVkRonBlf0CRNYc8Zttxdo=
-github.com/klauspost/compress v1.18.0/go.mod h1:2Pp+KzxcywXVXMr50+X0Q/Lsb43OQHYWRCY2AiWywWQ=
-github.com/konsorten/go-windows-terminal-sequences v1.0.1/go.mod h1:T0+1ngSBFLxvqU3pZ+m/2kptfBszLMUkC4ZK/EgS/cQ=
-github.com/kr/pretty v0.2.1/go.mod h1:ipq/a2n7PKx3OHsz4KJII5eveXtPO4qwEXGdVfWzfnI=
+github.com/klauspost/compress v1.19.1 h1:VsB4HPswih7mmZ8WleSFQ75c/Ui1M4trX5oAsJnhSlk=
+github.com/klauspost/compress v1.19.1/go.mod h1:cwPg85FWrGar70rWktvGQj8/hthj3wpl0PGDogxkrSQ=
 github.com/kr/pretty v0.3.1 h1:flRD4NNwYAUpkphVc1HcthR4KEIFJ65n8Mw5qdRn3LE=
 github.com/kr/pretty v0.3.1/go.mod h1:hoEshYVHaxMs3cyo3Yncou5ZscifuDolrwPKZanG3xk=
-github.com/kr/pty v1.1.1/go.mod h1:pFQYn66WHrOpPYNljwOMqo10TkYh1fy3cYio2l3bCsQ=
-github.com/kr/text v0.1.0/go.mod h1:4Jbv+DJW3UT/LiOwJeYQe1efqtUx/iVham/4vfdArNI=
 github.com/kr/text v0.2.0 h1:5Nx0Ya0ZqY2ygV366QzturHI13Jq95ApcVaJBhpS+AY=
 github.com/kr/text v0.2.0/go.mod h1:eLer722TekiGuMkidMxC/pM04lWEeraHUUmBw8l2grE=
 github.com/kylelemons/godebug v1.1.0 h1:RPNrshWIDI6G2gRW9EHilWtl7Z6Sb1BR0xunSBf0SNc=
 github.com/kylelemons/godebug v1.1.0/go.mod h1:9/0rRGxNHcop5bhtWyNeEfOS8JIWk580+fNqagV/RAw=
-github.com/magiconair/properties v1.8.9 h1:nWcCbLq1N2v/cpNsy5WvQ37Fb+YElfq20WJ/a8RkpQM=
-github.com/magiconair/properties v1.8.9/go.mod h1:Dhd985XPs7jluiymwWYZ0G4Z61jb3vdS329zhj2hYo0=
 github.com/mailru/easyjson v0.9.0 h1:PrnmzHw7262yW8sTBwxi1PdJA3Iw/EKBa8psRf7d9a4=
 github.com/mailru/easyjson v0.9.0/go.mod h1:1+xMtQp2MRNVL/V1bOzuP3aP8VNwRW55fQUto+XFtTU=
+github.com/maruel/natural v1.1.1 h1:Hja7XhhmvEFhcByqDoHz9QZbkWey+COd9xWfCfn1ioo=
+github.com/maruel/natural v1.1.1/go.mod h1:v+Rfd79xlw1AgVBjbO0BEQmptqb5HvL/k9GRHB7ZKEg=
 github.com/mattn/go-isatty v0.0.20 h1:xfD0iDuEKnDkl03q4limB+vH+GxLEtL/jb4xVJSWWEY=
 github.com/mattn/go-isatty v0.0.20/go.mod h1:W+V8PltTTMOvKvAeJH7IuucS94S2C6jfK/D7dTCTo3Y=
-github.com/mdlayher/ethtool v0.0.0-20210210192532-2b88debcdd43/go.mod h1:+t7E0lkKfbBsebllff1xdTmyJt8lH37niI6kwFk9OTo=
-github.com/mdlayher/ethtool v0.0.0-20211028163843-288d040e9d60/go.mod h1:aYbhishWc4Ai3I2U4Gaa2n3kHWSwzme6EsG/46HRQbE=
-github.com/mdlayher/genetlink v1.0.0/go.mod h1:0rJ0h4itni50A86M2kHcgS85ttZazNt7a8H2a2cw0Gc=
 github.com/mdlayher/genetlink v1.3.2 h1:KdrNKe+CTu+IbZnm/GVUMXSqBBLqcGpRDa0xkQy56gw=
 github.com/mdlayher/genetlink v1.3.2/go.mod h1:tcC3pkCrPUGIKKsCsp0B3AdaaKuHtaxoJRz3cc+528o=
 github.com/mdlayher/ndp v1.1.0 h1:QylGKGVtH60sKZUE88+IW5ila1Z/M9/OXhWdsVKuscs=
 github.com/mdlayher/ndp v1.1.0/go.mod h1:FmgESgemgjl38vuOIyAHWUUL6vQKA/pQNkvXdWsdQFM=
-github.com/mdlayher/netlink v0.0.0-20190409211403-11939a169225/go.mod h1:eQB3mZE4aiYnlUsyGGCOpPETfdQq4Jhsgf1fk3cwQaA=
-github.com/mdlayher/netlink v1.0.0/go.mod h1:KxeJAFOFLG6AjpyDkQ/iIhxygIUKD+vcwqcnu43w/+M=
-github.com/mdlayher/netlink v1.1.0/go.mod h1:H4WCitaheIsdF9yOYu8CFmCgQthAPIWZmcKp9uZHgmY=
-github.com/mdlayher/netlink v1.1.1/go.mod h1:WTYpFb/WTvlRJAyKhZL5/uy69TDDpHHu2VZmb2XgV7o=
-github.com/mdlayher/netlink v1.2.0/go.mod h1:kwVW1io0AZy9A1E2YYgaD4Cj+C+GPkU6klXCMzIJ9p8=
-github.com/mdlayher/netlink v1.2.1/go.mod h1:bacnNlfhqHqqLo4WsYeXSqfyXkInQ9JneWI68v1KwSU=
-github.com/mdlayher/netlink v1.2.2-0.20210123213345-5cc92139ae3e/go.mod h1:bacnNlfhqHqqLo4WsYeXSqfyXkInQ9JneWI68v1KwSU=
-github.com/mdlayher/netlink v1.3.0/go.mod h1:xK/BssKuwcRXHrtN04UBkwQ6dY9VviGGuriDdoPSWys=
-github.com/mdlayher/netlink v1.4.0/go.mod h1:dRJi5IABcZpBD2A3D0Mv/AiX8I9uDEu5oGkAVrekmf8=
-github.com/mdlayher/netlink v1.4.1/go.mod h1:e4/KuJ+s8UhfUpO9z00/fDZZmhSrs+oxyqAS9cNgn6Q=
-github.com/mdlayher/netlink v1.5.0/go.mod h1:1Kr8BBFxGyUyNmztC9WLOayqYVAd2wsgOZm18nqGuzQ=
-github.com/mdlayher/netlink v1.7.3-0.20250113171957-fbb4dce95f42 h1:A1Cq6Ysb0GM0tpKMbdCXCIfBclan4oHk1Jb+Hrejirg=
-github.com/mdlayher/netlink v1.7.3-0.20250113171957-fbb4dce95f42/go.mod h1:BB4YCPDOzfy7FniQ/lxuYQ3dgmM2cZumHbK8RpTjN2o=
+github.com/mdlayher/netlink v1.6.2/go.mod h1:O1HXX2sIWSMJ3Qn1BYZk1yZM+7iMki/uYGGiwGyq/iU=
+github.com/mdlayher/netlink v1.8.0 h1:e7XNIYJKD7hUct3Px04RuIGJbBxy1/c4nX7D5YyvvlM=
+github.com/mdlayher/netlink v1.8.0/go.mod h1:UhgKXUlDQhzb09DrCl2GuRNEglHmhYoWAHid9HK3594=
 github.com/mdlayher/packet v1.1.2 h1:3Up1NG6LZrsgDVn6X4L9Ge/iyRyxFEFD9o6Pr3Q1nQY=
 github.com/mdlayher/packet v1.1.2/go.mod h1:GEu1+n9sG5VtiRE4SydOmX5GTwyyYlteZiFU+x0kew4=
-github.com/mdlayher/socket v0.0.0-20210307095302-262dc9984e00/go.mod h1:GAFlyu4/XV68LkQKYzKhIo/WW7j3Zi0YRAz/BOoanUc=
-github.com/mdlayher/socket v0.0.0-20211007213009-516dcbdf0267/go.mod h1:nFZ1EtZYK8Gi/k6QNu7z7CgO20i/4ExeQswwWuPmG/g=
-github.com/mdlayher/socket v0.1.0/go.mod h1:mYV5YIZAfHh4dzDVzI8x8tWLWCliuX8Mon5Awbj+qDs=
+github.com/mdlayher/socket v0.2.3/go.mod h1:bz12/FozYNH/VbvC3q7TRIK/Y6dH1kCKsXaUeXi/FmY=
 github.com/mdlayher/socket v0.5.1 h1:VZaqt6RkGkt2OE9l3GcC6nZkqD3xKeQLyfleW/uBcos=
 github.com/mdlayher/socket v0.5.1/go.mod h1:TjPLHI1UgwEv5J1B5q0zTZq12A/6H7nKmtTanQE37IQ=
+github.com/mfridman/tparse v0.18.0 h1:wh6dzOKaIwkUGyKgOntDW4liXSo37qg5AXbIhkMV3vE=
+github.com/mfridman/tparse v0.18.0/go.mod h1:gEvqZTuCgEhPbYk/2lS3Kcxg1GmTxxU7kTC8DvP0i/A=
 github.com/mikioh/ipaddr v0.0.0-20190404000644-d465c8ab6721 h1:RlZweED6sbSArvlE924+mUcZuXKLBHA35U7LN621Bws=
 github.com/mikioh/ipaddr v0.0.0-20190404000644-d465c8ab6721/go.mod h1:Ickgr2WtCLZ2MDGd4Gr0geeCH5HybhRJbonOgQpvSxc=
-github.com/mitchellh/mapstructure v1.5.0 h1:jeMsZIYE/09sWLaz43PL7Gy6RuMjD2eJVyuac5Z2hdY=
-github.com/mitchellh/mapstructure v1.5.0/go.mod h1:bFUtVrKA4DC2yAKiSyO/QUcy7e+RRV2QTWOzhPopBRo=
 github.com/moby/docker-image-spec v1.3.1 h1:jMKff3w6PgbfSa69GfNg+zN/XLhfXJGnEx3Nl2EsFP0=
 github.com/moby/docker-image-spec v1.3.1/go.mod h1:eKmb5VW8vQEh/BAr2yvVNvuiJuY6UIocYsFu/DxxRpo=
 github.com/moby/sys/atomicwriter v0.1.0 h1:kw5D/EqkBwsBFi0ss9v1VG3wIkVhzGvLklJ+w3A14Sw=
@@ -209,83 +185,82 @@ github.com/moby/term v0.5.2/go.mod h1:d3djjFCrjnB+fl8NJux+EJzu0msscUP+f8it8hPkFL
 github.com/modern-go/concurrent v0.0.0-20180228061459-e0a39a4cb421/go.mod h1:6dJC0mAP4ikYIbvyc7fijjWJddQyLn8Ig3JB5CqoB9Q=
 github.com/modern-go/concurrent v0.0.0-20180306012644-bacd9c7ef1dd h1:TRLaZ9cD/w8PVh93nsPXa1VrQ6jlwL5oN8l14QlcNfg=
 github.com/modern-go/concurrent v0.0.0-20180306012644-bacd9c7ef1dd/go.mod h1:6dJC0mAP4ikYIbvyc7fijjWJddQyLn8Ig3JB5CqoB9Q=
-github.com/modern-go/reflect2 v1.0.2 h1:xBagoLtFs94CBntxluKeaWgTMpvLxC4ur3nMaC9Gz0M=
 github.com/modern-go/reflect2 v1.0.2/go.mod h1:yWuevngMOJpCy52FWWMvUC8ws7m/LJsjYzDa0/r8luk=
-github.com/morikuni/aec v1.0.0 h1:nP9CBfwrvYnBRgY6qfDQkygYDmYwOilePFkwzv4dU8A=
-github.com/morikuni/aec v1.0.0/go.mod h1:BbKIizmSmc5MMPqRYbxO4ZU0S0+P200+tUnFx7PXmsc=
+github.com/modern-go/reflect2 v1.0.3-0.20250322232337-35a7c28c31ee h1:W5t00kpgFdJifH4BDsTlE89Zl93FEloxaWZfGcifgq8=
+github.com/modern-go/reflect2 v1.0.3-0.20250322232337-35a7c28c31ee/go.mod h1:yWuevngMOJpCy52FWWMvUC8ws7m/LJsjYzDa0/r8luk=
+github.com/morikuni/aec v1.1.0 h1:vBBl0pUnvi/Je71dsRrhMBtreIqNMYErSAbEeb8jrXQ=
+github.com/morikuni/aec v1.1.0/go.mod h1:xDRgiq/iw5l+zkao76YTKzKttOp2cwPEne25HDkJnBw=
 github.com/munnerz/goautoneg v0.0.0-20191010083416-a7dc8b61c822 h1:C3w9PqII01/Oq1c1nUAm88MOHcQC9l5mIlSMApZMrHA=
 github.com/munnerz/goautoneg v0.0.0-20191010083416-a7dc8b61c822/go.mod h1:+n7T8mK8HuQTcFwEeznm/DIxMOiR9yIdICNftLE1DvQ=
-github.com/onsi/ginkgo/v2 v2.23.4 h1:ktYTpKJAVZnDT4VjxSbiBenUjmlL/5QkBEocaWXiQus=
-github.com/onsi/ginkgo/v2 v2.23.4/go.mod h1:Bt66ApGPBFzHyR+JO10Zbt0Gsp4uWxu5mIOTusL46e8=
-github.com/onsi/gomega v1.37.0 h1:CdEG8g0S133B4OswTDC/5XPSzE1OeP29QOioj2PID2Y=
-github.com/onsi/gomega v1.37.0/go.mod h1:8D9+Txp43QWKhM24yyOBEdpkzN8FvJyAwecBgsU4KU0=
+github.com/onsi/ginkgo/v2 v2.32.1 h1:6tlvcDm/3sE8lGJbZ4+d4mO3RLy24/tQWOFzVSQNIfw=
+github.com/onsi/ginkgo/v2 v2.32.1/go.mod h1:+aXOY+vzZ5mu2iI2HpTZUPmM//oQfsNFX6gU9kNcA44=
+github.com/onsi/gomega v1.42.1 h1:iN1rCUX+44NZ1Dc97MPoeFYbFR0vh8zxoxMFwKdyZ6I=
+github.com/onsi/gomega v1.42.1/go.mod h1:REff/hsDsodHoKlWsP2mAPhu1+5/6hVYNf9rIEBpeSg=
 github.com/opencontainers/go-digest v1.0.0 h1:apOUWs51W5PlhuyGyz9FCeeBIOUDA/6nW8Oi/yOhh5U=
 github.com/opencontainers/go-digest v1.0.0/go.mod h1:0JzlMkj0TRzQZfJkVvzbP0HBR3IKzErnv2BNG4W4MAM=
 github.com/opencontainers/image-spec v1.1.1 h1:y0fUlFfIZhPF1W537XOLg0/fcx6zcHCJwooC2xJA040=
 github.com/opencontainers/image-spec v1.1.1/go.mod h1:qpqAh3Dmcf36wStyyWU+kCeDgrGnAve2nCC8+7h8Q0M=
-github.com/osrg/gobgp/v3 v3.37.0 h1:+ObuOdvj7G7nxrT0fKFta+EAupdWf/q1WzbXydr8IOY=
-github.com/osrg/gobgp/v3 v3.37.0/go.mod h1:kVHVFy1/fyZHJ8P32+ctvPeJogn9qKwa1YCeMRXXrP0=
-github.com/pbnjay/memory v0.0.0-20210728143218-7b4eea64cf58 h1:onHthvaw9LFnH4t2DcNVpwGmV9E1BkGknEliJkfwQj0=
-github.com/pbnjay/memory v0.0.0-20210728143218-7b4eea64cf58/go.mod h1:DXv8WO4yhMYhSNPKjeNKa5WY9YCIEBRbNzFFPJbWO6Y=
+github.com/orcaman/concurrent-map/v2 v2.0.1 h1:jOJ5Pg2w1oeB6PeDurIYf6k9PQ+aTITr/6lP/L/zp6c=
+github.com/orcaman/concurrent-map/v2 v2.0.1/go.mod h1:9Eq3TG2oBe5FirmYWQfYO5iH1q0Jv47PLaNK++uCdOM=
+github.com/osrg/gobgp/v4 v4.8.0 h1:CoprkOZ2nsH8aTHij6xWj18QOrsBwG9D00l1PGw7394=
+github.com/osrg/gobgp/v4 v4.8.0/go.mod h1:bJbFm7T2nRANggShfl3I9h0UpPCzu4uAY5J/6dTdRvs=
 github.com/pelletier/go-toml v1.9.5 h1:4yBQzkHv+7BHq2PQUZF3Mx0IYxG7LsP222s7Agd3ve8=
 github.com/pelletier/go-toml v1.9.5/go.mod h1:u1nR/EPcESfeI/szUZKdtJ0xRNbUoANCkoOuaOx1Y+c=
 github.com/pelletier/go-toml/v2 v2.2.3 h1:YmeHyLY8mFWbdkNWwpr+qIL2bEqT0o95WSdkNHvL12M=
 github.com/pelletier/go-toml/v2 v2.2.3/go.mod h1:MfCQTFTvCcUyyvvwm1+G6H/jORL20Xlb6rzQu9GuUkc=
 github.com/pierrec/lz4/v4 v4.1.22 h1:cKFw6uJDK+/gfw5BcDL0JL5aBsAFdsIT18eRtLj7VIU=
 github.com/pierrec/lz4/v4 v4.1.22/go.mod h1:gZWDp/Ze/IJXGXf23ltt2EXimqmTUXEy0GFuRQyBid4=
-github.com/pkg/errors v0.8.1/go.mod h1:bwawxfHBFNV+L2hUp1rHADufV3IMtnDRdf1r5NINEl0=
 github.com/pkg/errors v0.9.1 h1:FEBLx1zS214owpjy7qsBeixbURkuhQAwrK5UwLGTwt4=
 github.com/pkg/errors v0.9.1/go.mod h1:bwawxfHBFNV+L2hUp1rHADufV3IMtnDRdf1r5NINEl0=
 github.com/pmezard/go-difflib v1.0.0/go.mod h1:iKH77koFhYxTK1pcRnkKkqfTogsbg7gZNVY4sRDYZ/4=
 github.com/pmezard/go-difflib v1.0.1-0.20181226105442-5d4384ee4fb2 h1:Jamvg5psRIccs7FGNTlIRMkT8wgtp5eCXdBlqhYGL6U=
 github.com/pmezard/go-difflib v1.0.1-0.20181226105442-5d4384ee4fb2/go.mod h1:iKH77koFhYxTK1pcRnkKkqfTogsbg7gZNVY4sRDYZ/4=
-github.com/prashantv/gostub v1.1.0 h1:BTyx3RfQjRHnUWaGF9oQos79AlQ5k8WNktv7VGvVH4g=
-github.com/prashantv/gostub v1.1.0/go.mod h1:A5zLQHz7ieHGG7is6LLXLz7I8+3LZzsrV0P1IAHhP5U=
-github.com/prometheus/client_golang v1.22.0 h1:rb93p9lokFEsctTys46VnV1kLCDpVZ0a/Y92Vm0Zc6Q=
-github.com/prometheus/client_golang v1.22.0/go.mod h1:R7ljNsLXhuQXYZYtw6GAE9AZg8Y7vEW5scdCXrWRXC0=
-github.com/prometheus/client_model v0.6.1 h1:ZKSh/rekM+n3CeS952MLRAdFwIKqeY8b62p8ais2e9E=
-github.com/prometheus/client_model v0.6.1/go.mod h1:OrxVMOVHjw3lKMa8+x6HeMGkHMQyHDk9E3jmP2AmGiY=
-github.com/prometheus/common v0.62.0 h1:xasJaQlnWAeyHdUBeGjXmutelfJHWMRr+Fg4QszZ2Io=
-github.com/prometheus/common v0.62.0/go.mod h1:vyBcEuLSvWos9B1+CyL7JZ2up+uFzXhkqml0W5zIY1I=
-github.com/prometheus/procfs v0.15.1 h1:YagwOFzUgYfKKHX6Dr+sHT7km/hxC76UB0learggepc=
-github.com/prometheus/procfs v0.15.1/go.mod h1:fB45yRUv8NstnjriLhBQLuOUt+WW4BsoGhij/e3PBqk=
-github.com/remyoudompheng/bigfft v0.0.0-20230129092748-24d4a6f8daec h1:W09IVJc94icq4NjY3clb7Lk8O1qJ8BdBEF8z0ibU0rE=
-github.com/remyoudompheng/bigfft v0.0.0-20230129092748-24d4a6f8daec/go.mod h1:qqbHyh8v60DhA7CoWK5oRCqLrMHRGoxYCSS9EjAz6Eo=
-github.com/rogpeppe/go-internal v1.13.1 h1:KvO1DLK/DRN07sQ1LQKScxyZJuNnedQ5/wKSR38lUII=
-github.com/rogpeppe/go-internal v1.13.1/go.mod h1:uMEvuHeurkdAXX61udpOXGD/AzZDWNMNyH2VO9fmH0o=
+github.com/prometheus/client_golang v1.24.1 h1:JnJkREXzWxUdCuPFpIWZiPispT9xVV59uiuyR2bPlnU=
+github.com/prometheus/client_golang v1.24.1/go.mod h1:F+oSRECHg4sse5ucfYpYDeIv/hu68Zo0uoHKetWnzcE=
+github.com/prometheus/client_model v0.6.2 h1:oBsgwpGs7iVziMvrGhE53c/GrLUsZdHnqNwqPLxwZyk=
+github.com/prometheus/client_model v0.6.2/go.mod h1:y3m2F6Gdpfy6Ut/GBsUqTWZqCUvMVzSfMLjcu6wAwpE=
+github.com/prometheus/common v0.70.1 h1:1HvjP4D5oL3t8RsPlwxA9onvvStjtIHYE5XuuwOi/PY=
+github.com/prometheus/common v0.70.1/go.mod h1:VdFUQDMZK3VLkurFUVhia6uys/0suUp86TJz5qbJRhc=
+github.com/prometheus/procfs v0.21.1 h1:GljZCt+zSTS+NZq88cyQ1LjZ+RCHp3uVuabBWA5+OJI=
+github.com/prometheus/procfs v0.21.1/go.mod h1:aB55Cww9pdSJVHk0hUf0inxWyyjPogFIjmHKYgMKmtY=
+github.com/rogpeppe/go-internal v1.14.1 h1:UQB4HGPB6osV0SQTLymcB4TgvyWu6ZyliaW0tI/otEQ=
+github.com/rogpeppe/go-internal v1.14.1/go.mod h1:MaRKkUm5W0goXpeCfT7UZI6fk/L7L7so1lCWt35ZSgc=
 github.com/russross/blackfriday/v2 v2.1.0/go.mod h1:+Rmxgy9KzJVeS9/2gXHxylqXiyQDYRxCVz55jmeOWTM=
-github.com/sagikazarmark/locafero v0.6.0 h1:ON7AQg37yzcRPU69mt7gwhFEBwxI6P9T4Qu3N51bwOk=
-github.com/sagikazarmark/locafero v0.6.0/go.mod h1:77OmuIc6VTraTXKXIs/uvUxKGUXjE1GbemJYHqdNjX0=
-github.com/sagikazarmark/slog-shim v0.1.0 h1:diDBnUNK9N/354PgrxMywXnAwEr1QZcOr6gto+ugjYE=
-github.com/sagikazarmark/slog-shim v0.1.0/go.mod h1:SrcSrq8aKtyuqEI1uvTDTK1arOWRIczQRv+GVI1AkeQ=
-github.com/sirupsen/logrus v1.4.1/go.mod h1:ni0Sbl8bgC9z8RoU9G6nDWqqs/fq4eDPysMBDgk/93Q=
-github.com/sirupsen/logrus v1.9.3 h1:dueUQJ1C2q9oE3F7wvmSGAaVtTmUizReu6fjN8uqzbQ=
-github.com/sirupsen/logrus v1.9.3/go.mod h1:naHLuLoDiP4jHNo9R0sCBMtWGeIprob74mVsIT4qYEQ=
+github.com/sagikazarmark/locafero v0.7.0 h1:5MqpDsTGNDhY8sGp0Aowyf0qKsPrhewaLSsFaodPcyo=
+github.com/sagikazarmark/locafero v0.7.0/go.mod h1:2za3Cg5rMaTMoG/2Ulr9AwtFaIppKXTRYnozin4aB5k=
+github.com/segmentio/fasthash v1.0.3 h1:EI9+KE1EwvMLBWwjpRDc+fEM+prwxDYbslddQGtrmhM=
+github.com/segmentio/fasthash v1.0.3/go.mod h1:waKX8l2N8yckOgmSsXJi7x1ZfdKZ4x7KRMzBtS3oedY=
+github.com/sirupsen/logrus v1.9.4 h1:TsZE7l11zFCLZnZ+teH4Umoq5BhEIfIzfRDZ1Uzql2w=
+github.com/sirupsen/logrus v1.9.4/go.mod h1:ftWc9WdOfJ0a92nsE2jF5u5ZwH8Bv2zdeOC42RjbV2g=
 github.com/sourcegraph/conc v0.3.0 h1:OQTbbt6P72L20UqAkXXuLOj79LfEanQ+YQFNpLA9ySo=
 github.com/sourcegraph/conc v0.3.0/go.mod h1:Sdozi7LEKbFPqYX2/J+iBAM6HpqSLTASQIKqDmF7Mt0=
-github.com/spf13/afero v1.11.0 h1:WJQKhtpdm3v2IzqG8VMqrr6Rf3UYpEF239Jy9wNepM8=
-github.com/spf13/afero v1.11.0/go.mod h1:GH9Y3pIexgf1MTIWtNGyogA5MwRIDXGUr+hbWNoBjkY=
+github.com/spf13/afero v1.12.0 h1:UcOPyRBYczmFn6yvphxkn9ZEOY65cpwGKb5mL36mrqs=
+github.com/spf13/afero v1.12.0/go.mod h1:ZTlWwG4/ahT8W7T0WQ5uYmjI9duaLQGy3Q2OAl4sk/4=
 github.com/spf13/cast v1.7.1 h1:cuNEagBQEHWN1FnbGEjCXL2szYEXqfJPbP2HNUaca9Y=
 github.com/spf13/cast v1.7.1/go.mod h1:ancEpBxwJDODSW/UG4rDrAqiKolqNNh2DX3mk86cAdo=
-github.com/spf13/cobra v1.9.1 h1:CXSaggrXdbHK9CF+8ywj8Amf7PBRmPCOJugH954Nnlo=
-github.com/spf13/cobra v1.9.1/go.mod h1:nDyEzZ8ogv936Cinf6g1RU9MRY64Ir93oCnqb9wxYW0=
-github.com/spf13/pflag v1.0.6 h1:jFzHGLGAlb3ruxLB8MhbI6A8+AQX/2eW4qeyNZXNp2o=
-github.com/spf13/pflag v1.0.6/go.mod h1:McXfInJRrz4CZXVZOBLb0bTZqETkiAhM9Iw0y3An2Bg=
-github.com/spf13/viper v1.19.0 h1:RWq5SEjt8o25SROyN3z2OrDB9l7RPd3lwTWU8EcEdcI=
-github.com/spf13/viper v1.19.0/go.mod h1:GQUN9bilAbhU/jgc1bKs99f/suXKeUMct8Adx5+Ntkg=
+github.com/spf13/cobra v1.10.2 h1:DMTTonx5m65Ic0GOoRY2c16WCbHxOOw6xxezuLaBpcU=
+github.com/spf13/cobra v1.10.2/go.mod h1:7C1pvHqHw5A4vrJfjNwvOdzYu0Gml16OCs2GRiTUUS4=
+github.com/spf13/pflag v1.0.9/go.mod h1:McXfInJRrz4CZXVZOBLb0bTZqETkiAhM9Iw0y3An2Bg=
+github.com/spf13/pflag v1.0.10 h1:4EBh2KAYBwaONj6b2Ye1GiHfwjqyROoF4RwYO+vPwFk=
+github.com/spf13/pflag v1.0.10/go.mod h1:McXfInJRrz4CZXVZOBLb0bTZqETkiAhM9Iw0y3An2Bg=
+github.com/spf13/viper v1.20.1 h1:ZMi+z/lvLyPSCoNtFCpqjy0S4kPbirhpTMwl8BkW9X4=
+github.com/spf13/viper v1.20.1/go.mod h1:P9Mdzt1zoHIG8m2eZQinpiBjo6kCmZSKBClNNqjJvu4=
 github.com/stretchr/objx v0.1.0/go.mod h1:HFkY916IF+rwdDfMAkV7OtwuqBVzrE8GR6GFx+wExME=
-github.com/stretchr/objx v0.1.1/go.mod h1:HFkY916IF+rwdDfMAkV7OtwuqBVzrE8GR6GFx+wExME=
 github.com/stretchr/objx v0.5.2 h1:xuMeJ0Sdp5ZMRXx/aWO6RZxdr3beISkG5/G/aIRr3pY=
 github.com/stretchr/objx v0.5.2/go.mod h1:FRsXN1f5AsAjCGJKqEizvkpNtU+EGNCLh3NxZ/8L+MA=
-github.com/stretchr/testify v1.2.2/go.mod h1:a8OnRcib4nhh0OaRAV+Yts87kKdq0PP7pXfy6kDkUVs=
 github.com/stretchr/testify v1.3.0/go.mod h1:M5WIy9Dh21IEIfnGCwXGc5bZfKNJtfHm1UVUgZn+9EI=
-github.com/stretchr/testify v1.7.0/go.mod h1:6Fq8oRcR53rry900zMqJjRRixrwX3KX962/h/Wwjteg=
-github.com/stretchr/testify v1.10.0 h1:Xv5erBjTwe/5IxqUQTdXv5kgmIvbHo3QQyRwhJsOfJA=
-github.com/stretchr/testify v1.10.0/go.mod h1:r2ic/lqez/lEtzL7wO/rwa5dbSLXVDPFyf8C91i36aY=
+github.com/stretchr/testify v1.11.1 h1:7s2iGBzp5EwR7/aIZr8ao5+dra3wiQyKjjFuvgVKu7U=
+github.com/stretchr/testify v1.11.1/go.mod h1:wZwfW3scLgRK+23gO65QZefKpKQRnfz6sD981Nm4B6U=
 github.com/subosito/gotenv v1.6.0 h1:9NlTDc1FTs4qu0DDq7AEtTPNw6SVm7uBMsUCUjABIf8=
 github.com/subosito/gotenv v1.6.0/go.mod h1:Dk4QP5c2W3ibzajGcXpNraDfq2IrhjMIvMSWPKKo0FU=
-github.com/tj/go-spin v1.1.0 h1:lhdWZsvImxvZ3q1C5OIB7d72DuOwP4O2NdBg9PyzNds=
-github.com/tj/go-spin v1.1.0/go.mod h1:Mg1mzmePZm4dva8Qz60H2lHwmJ2loum4VIrLgVnKwh4=
+github.com/tidwall/gjson v1.18.0 h1:FIDeeyB800efLX89e5a8Y0BNH+LOngJyGrIWxG2FKQY=
+github.com/tidwall/gjson v1.18.0/go.mod h1:/wbyibRr2FHMks5tjHJ5F8dMZh3AcwJEMf5vlfC0lxk=
+github.com/tidwall/match v1.1.1 h1:+Ho715JplO36QYgwN9PGYNhgZvoUSc9X2c80KVTi+GA=
+github.com/tidwall/match v1.1.1/go.mod h1:eRSPERbgtNPcGhD8UCthc6PmLEQXEWd3PRB5JTxsfmM=
+github.com/tidwall/pretty v1.2.1 h1:qjsOFOWWQl+N3RsoF5/ssm1pHmJJwhjlSbZ51I6wMl4=
+github.com/tidwall/pretty v1.2.1/go.mod h1:ITEVvHYasfjBbM0u2Pg8T2nJnzm8xPwvNhhsoaGGjNU=
+github.com/tidwall/sjson v1.2.5 h1:kLy8mja+1c9jlljvWTlSazM7cKDRfJuR/bOJhcY5NcY=
+github.com/tidwall/sjson v1.2.5/go.mod h1:Fvgq9kS/6ociJEDnK0Fk1cpYF4FIW6ZF7LAe+6jwd28=
 github.com/u-root/uio v0.0.0-20240224005618-d2acac8f3701 h1:pyC9PaHYZFgEKFdlp3G8RaCKgVpHZnecvArXvPXcFkM=
 github.com/u-root/uio v0.0.0-20240224005618-d2acac8f3701/go.mod h1:P3a5rG4X7tI17Nn3aOIAYr5HbIMukwXG0urG0WuL8OA=
 github.com/valyala/bytebufferpool v1.0.0 h1:GqA5TC/0021Y/b9FG4Oi9Mr3q7XYx6KllzawFIhcdPw=
@@ -296,214 +271,129 @@ github.com/vishvananda/netns v0.0.5 h1:DfiHV+j8bA32MFM7bfEunvT8IAqQ/NzSJHtcmW5zd
 github.com/vishvananda/netns v0.0.5/go.mod h1:SpkAiCQRtJ6TvvxPnOSyH3BMl6unz3xZlaprSwhNNJM=
 github.com/x448/float16 v0.8.4 h1:qLwI1I70+NjRFUR3zs1JPUCgaCXSh3SW62uAKT1mSBM=
 github.com/x448/float16 v0.8.4/go.mod h1:14CWIYCyZA/cWjXOioeEpHeN/83MdbZDRQHoFcYsOfg=
-github.com/xlab/c-for-go v1.3.0 h1:WDc+/E59g1OcZ4dYB4K9V4Doh4NnVBMYxaZFhNi77Lg=
-github.com/xlab/c-for-go v1.3.0/go.mod h1:0qVnGIfoNg7c5gABptM38Y7nnXgAMo5KYWVQPycu63Y=
-github.com/xlab/pkgconfig v0.0.0-20170226114623-cea12a0fd245 h1:Sw125DKxZhPUI4JLlWugkzsrlB50jR9v2khiD9FxuSo=
-github.com/xlab/pkgconfig v0.0.0-20170226114623-cea12a0fd245/go.mod h1:C+diUUz7pxhNY6KAoLgrTYARGWnt82zWTylZlxT92vk=
 github.com/xo/terminfo v0.0.0-20220910002029-abceb7e1c41e h1:JVG44RsyaB9T2KIHavMF/ppJZNG9ZpyihvCd0w101no=
 github.com/xo/terminfo v0.0.0-20220910002029-abceb7e1c41e/go.mod h1:RbqR21r5mrJuqunuUZ/Dhy/avygyECGrLceyNeo4LiM=
-github.com/yuin/goldmark v1.1.27/go.mod h1:3hX8gzYuyVAZsxl0MRgGTJEmQBFcNTphYh9decYSb74=
-github.com/yuin/goldmark v1.2.1/go.mod h1:3hX8gzYuyVAZsxl0MRgGTJEmQBFcNTphYh9decYSb74=
-github.com/yuin/goldmark v1.4.0/go.mod h1:mwnBkeHKe2W/ZEtQ+71ViKU8L12m81fl3OWwC1Zlc8k=
-go.etcd.io/etcd/api/v3 v3.6.1 h1:yJ9WlDih9HT457QPuHt/TH/XtsdN2tubyxyQHSHPsEo=
-go.etcd.io/etcd/api/v3 v3.6.1/go.mod h1:lnfuqoGsXMlZdTJlact3IB56o3bWp1DIlXPIGKRArto=
-go.etcd.io/etcd/client/pkg/v3 v3.6.1 h1:CxDVv8ggphmamrXM4Of8aCC8QHzDM4tGcVr9p2BSoGk=
-go.etcd.io/etcd/client/pkg/v3 v3.6.1/go.mod h1:aTkCp+6ixcVTZmrJGa7/Mc5nMNs59PEgBbq+HCmWyMc=
-go.etcd.io/etcd/client/v3 v3.6.1 h1:KelkcizJGsskUXlsxjVrSmINvMMga0VWwFF0tSPGEP0=
-go.etcd.io/etcd/client/v3 v3.6.1/go.mod h1:fCbPUdjWNLfx1A6ATo9syUmFVxqHH9bCnPLBZmnLmMY=
-go.opentelemetry.io/auto/sdk v1.1.0 h1:cH53jehLUN6UFLY71z+NDOiNJqDdPRaXzTel0sJySYA=
-go.opentelemetry.io/auto/sdk v1.1.0/go.mod h1:3wSPjt5PWp2RhlCcmmOial7AvC4DQqZb7a7wCow3W8A=
-go.opentelemetry.io/contrib/instrumentation/net/http/otelhttp v0.49.0 h1:jq9TW8u3so/bN+JPT166wjOI6/vQPF6Xe7nMNIltagk=
-go.opentelemetry.io/contrib/instrumentation/net/http/otelhttp v0.49.0/go.mod h1:p8pYQP+m5XfbZm9fxtSKAbM6oIllS7s2AfxrChvc7iw=
-go.opentelemetry.io/otel v1.37.0 h1:9zhNfelUvx0KBfu/gb+ZgeAfAgtWrfHJZcAqFC228wQ=
-go.opentelemetry.io/otel v1.37.0/go.mod h1:ehE/umFRLnuLa/vSccNq9oS1ErUlkkK71gMcN34UG8I=
-go.opentelemetry.io/otel/exporters/otlp/otlptrace v1.37.0 h1:Ahq7pZmv87yiyn3jeFz/LekZmPLLdKejuO3NcK9MssM=
-go.opentelemetry.io/otel/exporters/otlp/otlptrace v1.37.0/go.mod h1:MJTqhM0im3mRLw1i8uGHnCvUEeS7VwRyxlLC78PA18M=
-go.opentelemetry.io/otel/exporters/otlp/otlptrace/otlptracehttp v1.37.0 h1:bDMKF3RUSxshZ5OjOTi8rsHGaPKsAt76FaqgvIUySLc=
-go.opentelemetry.io/otel/exporters/otlp/otlptrace/otlptracehttp v1.37.0/go.mod h1:dDT67G/IkA46Mr2l9Uj7HsQVwsjASyV9SjGofsiUZDA=
-go.opentelemetry.io/otel/metric v1.37.0 h1:mvwbQS5m0tbmqML4NqK+e3aDiO02vsf/WgbsdpcPoZE=
-go.opentelemetry.io/otel/metric v1.37.0/go.mod h1:04wGrZurHYKOc+RKeye86GwKiTb9FKm1WHtO+4EVr2E=
-go.opentelemetry.io/otel/sdk v1.37.0 h1:ItB0QUqnjesGRvNcmAcU0LyvkVyGJ2xftD29bWdDvKI=
-go.opentelemetry.io/otel/sdk v1.37.0/go.mod h1:VredYzxUvuo2q3WRcDnKDjbdvmO0sCzOvVAiY+yUkAg=
-go.opentelemetry.io/otel/sdk/metric v1.35.0 h1:1RriWBmCKgkeHEhM7a2uMjMUfP7MsOF5JpUCaEqEI9o=
-go.opentelemetry.io/otel/sdk/metric v1.35.0/go.mod h1:is6XYCUMpcKi+ZsOvfluY5YstFnhW0BidkR+gL+qN+w=
-go.opentelemetry.io/otel/trace v1.37.0 h1:HLdcFNbRQBE2imdSEgm/kwqmQj1Or1l/7bW6mxVK7z4=
-go.opentelemetry.io/otel/trace v1.37.0/go.mod h1:TlgrlQ+PtQO5XFerSPUYG0JSgGyryXewPGyayAWSBS0=
-go.opentelemetry.io/proto/otlp v1.7.0 h1:jX1VolD6nHuFzOYso2E73H85i92Mv8JQYk0K9vz09os=
-go.opentelemetry.io/proto/otlp v1.7.0/go.mod h1:fSKjH6YJ7HDlwzltzyMj036AJ3ejJLCgCSHGj4efDDo=
-go.uber.org/automaxprocs v1.6.0 h1:O3y2/QNTOdbF+e/dpXNNW7Rx2hZ4sTIPyybbxyNqTUs=
-go.uber.org/automaxprocs v1.6.0/go.mod h1:ifeIMSnPZuznNm6jmdzmU3/bfk01Fe2fotchwEFJ8r8=
+go.etcd.io/etcd/api/v3 v3.7.1 h1:KJG0/DcWGfe3Y1otDf/fsBf0TSSgpxZ5RO/L8SFt73E=
+go.etcd.io/etcd/api/v3 v3.7.1/go.mod h1:8bXIpCMeV7E3/XL0Ix123ATn3dB+0V7d9zklHbB0m78=
+go.etcd.io/etcd/client/pkg/v3 v3.7.1 h1:rKYsj3pRkR0eK3yjT3XOgrhqfmIfj9pzNgxjh7mfFv4=
+go.etcd.io/etcd/client/pkg/v3 v3.7.1/go.mod h1:cnzZGIUzSfjEwLC6UBVsSXlEK1eepS/JUD7wE6PLRT0=
+go.etcd.io/etcd/client/v3 v3.7.1 h1:0PEMMC0KuZmVIN+RAbdqfkZ45pYTgKVtmBEbRCvZFUg=
+go.etcd.io/etcd/client/v3 v3.7.1/go.mod h1:ffNqALa8tRCYhYo1F9oR489y23K39Gz+BSR3ApAGYq0=
+go.opentelemetry.io/auto/sdk v1.2.1 h1:jXsnJ4Lmnqd11kwkBV2LgLoFMZKizbCi5fNZ/ipaZ64=
+go.opentelemetry.io/auto/sdk v1.2.1/go.mod h1:KRTj+aOaElaLi+wW1kO/DZRXwkF4C5xPbEe3ZiIhN7Y=
+go.opentelemetry.io/contrib/instrumentation/net/http/otelhttp v0.61.0 h1:F7Jx+6hwnZ41NSFTO5q4LYDtJRXBf2PD0rNBkeB/lus=
+go.opentelemetry.io/contrib/instrumentation/net/http/otelhttp v0.61.0/go.mod h1:UHB22Z8QsdRDrnAtX4PntOl36ajSxcdUMt1sF7Y6E7Q=
+go.opentelemetry.io/otel v1.44.0 h1:JjwHmHpA4iZ3wBxluu2fbbE7j4kqlE8jXyAyPXH7HqU=
+go.opentelemetry.io/otel v1.44.0/go.mod h1:BMgjTHL9WPRlRjL2oZCBTL4whCGtXch2H4BhOPIAyYc=
+go.opentelemetry.io/otel/exporters/otlp/otlptrace v1.43.0 h1:88Y4s2C8oTui1LGM6bTWkw0ICGcOLCAI5l6zsD1j20k=
+go.opentelemetry.io/otel/exporters/otlp/otlptrace v1.43.0/go.mod h1:Vl1/iaggsuRlrHf/hfPJPvVag77kKyvrLeD10kpMl+A=
+go.opentelemetry.io/otel/exporters/otlp/otlptrace/otlptracehttp v1.43.0 h1:3iZJKlCZufyRzPzlQhUIWVmfltrXuGyfjREgGP3UUjc=
+go.opentelemetry.io/otel/exporters/otlp/otlptrace/otlptracehttp v1.43.0/go.mod h1:/G+nUPfhq2e+qiXMGxMwumDrP5jtzU+mWN7/sjT2rak=
+go.opentelemetry.io/otel/metric v1.44.0 h1:1w0gILTcHdr3YI+ixLyjemwrVnsMURbTZFrSYCdDdmc=
+go.opentelemetry.io/otel/metric v1.44.0/go.mod h1:8O7hanEPBNgEMmybD3s2VBKcgWOCsA6tzHBPODAiquo=
+go.opentelemetry.io/otel/sdk v1.44.0 h1:nHYwb9lK+fJPU/dnT6s7W7Z8itMWyqrnVfbheVYrZ58=
+go.opentelemetry.io/otel/sdk v1.44.0/go.mod h1:Osuydd3Se74nqjAKxid74N5eC+jfEqfTegHRnq58oK0=
+go.opentelemetry.io/otel/sdk/metric v1.44.0 h1:3LlKgI+VjbVsjNRFZJZAJ30WjXC5VkNRks6si09iEfI=
+go.opentelemetry.io/otel/sdk/metric v1.44.0/go.mod h1:5B5pMARnXxKhltooO4xUuCBorl65a4EpnTalObqOigA=
+go.opentelemetry.io/otel/trace v1.44.0 h1:jxF5CsGYCe74MCRx2X4g7WsY/VBKRqqpNvXlX/6gtIk=
+go.opentelemetry.io/otel/trace v1.44.0/go.mod h1:oLl1jrMQAVo6v3GAggN+1VH9VIz9iUSvW53sW1Q8PIE=
+go.opentelemetry.io/proto/otlp v1.10.0 h1:IQRWgT5srOCYfiWnpqUYz9CVmbO8bFmKcwYxpuCSL2g=
+go.opentelemetry.io/proto/otlp v1.10.0/go.mod h1:/CV4QoCR/S9yaPj8utp3lvQPoqMtxXdzn7ozvvozVqk=
 go.uber.org/goleak v1.3.0 h1:2K3zAYmnTNqV73imy9J1T3WC+gmCePx2hEGkimedGto=
 go.uber.org/goleak v1.3.0/go.mod h1:CoHD4mav9JJNrW/WLlf7HGZPjdw8EucARQHekz1X6bE=
 go.uber.org/multierr v1.11.0 h1:blXXJkSxSSfBVBlC76pxqeO+LN3aDfLQo+309xJstO0=
 go.uber.org/multierr v1.11.0/go.mod h1:20+QtiLqy0Nd6FdQB9TLXag12DsQkrbs3htMFfDN80Y=
-go.uber.org/zap v1.27.0 h1:aJMhYGrd5QSmlpLMr2MftRKl7t8J8PTZPA732ud/XR8=
-go.uber.org/zap v1.27.0/go.mod h1:GB2qFLM7cTU87MWRP2mPIjqfIDnGu+VIO4V/SdhGo2E=
+go.uber.org/zap v1.28.0 h1:IZzaP1Fv73/T/pBMLk4VutPl36uNC+OSUh3JLG3FIjo=
+go.uber.org/zap v1.28.0/go.mod h1:rDLpOi171uODNm/mxFcuYWxDsqWSAVkFdX4XojSKg/Q=
+go.yaml.in/yaml/v2 v2.4.4 h1:tuyd0P+2Ont/d6e2rl3be67goVK4R6deVxCUX5vyPaQ=
+go.yaml.in/yaml/v2 v2.4.4/go.mod h1:gMZqIpDtDqOfM0uNfy0SkpRhvUryYH0Z6wdMYcacYXQ=
+go.yaml.in/yaml/v3 v3.0.4 h1:tfq32ie2Jv2UxXFdLJdh3jXuOzWiL1fo0bu/FbuKpbc=
+go.yaml.in/yaml/v3 v3.0.4/go.mod h1:DhzuOOF2ATzADvBadXxruRBLzYTpT36CKvDb3+aBEFg=
 golang.org/x/crypto v0.0.0-20190308221718-c2843e01d9a2/go.mod h1:djNgcEr1/C05ACkg1iLfiJU5Ep61QUkGW8qpdssI0+w=
-golang.org/x/crypto v0.0.0-20191011191535-87dc89f01550/go.mod h1:yigFU9vqHzYiE8UmvKecakEJjdnWj3jj499lnFckfCI=
-golang.org/x/crypto v0.0.0-20200622213623-75b288015ac9/go.mod h1:LzIPMQfyMNhhGPhUkYOs5KpL4U8rLKemX1yGLhDgUto=
-golang.org/x/crypto v0.39.0 h1:SHs+kF4LP+f+p14esP5jAoDpHU8Gu/v9lFRK6IT5imM=
-golang.org/x/crypto v0.39.0/go.mod h1:L+Xg3Wf6HoL4Bn4238Z6ft6KfEpN0tJGo53AAPC632U=
+golang.org/x/crypto v0.54.0 h1:YLIA59K4fiNzHzjnZt2tUJQjQtUWfWbeHBqKtk3eScw=
+golang.org/x/crypto v0.54.0/go.mod h1:KWL8ny2AZdGR2cWmzeHrp2azQPGogOv+HeQaVEXC2dk=
 golang.org/x/exp v0.0.0-20250103183323-7d7fa50e5329 h1:9kj3STMvgqy3YA4VQXBrN7925ICMxD5wzMRcgA30588=
 golang.org/x/exp v0.0.0-20250103183323-7d7fa50e5329/go.mod h1:qj5a5QZpwLU2NLQudwIN5koi3beDhSAlJwa67PuM98c=
-golang.org/x/mod v0.2.0/go.mod h1:s0Qsj1ACt9ePp/hMypM3fl4fZqREWJwdYDEqhRiZZUA=
-golang.org/x/mod v0.3.0/go.mod h1:s0Qsj1ACt9ePp/hMypM3fl4fZqREWJwdYDEqhRiZZUA=
-golang.org/x/mod v0.4.2/go.mod h1:s0Qsj1ACt9ePp/hMypM3fl4fZqREWJwdYDEqhRiZZUA=
-golang.org/x/mod v0.5.1/go.mod h1:5OXOZSfqPIIbmVBIIKWRFfZjPR0E5r58TLhUjH0a2Ro=
-golang.org/x/mod v0.25.0 h1:n7a+ZbQKQA/Ysbyb0/6IbB1H/X41mKgbhfv7AfG/44w=
-golang.org/x/mod v0.25.0/go.mod h1:IXM97Txy2VM4PJ3gI61r1YEk/gAj6zAHN3AdZt6S9Ww=
-golang.org/x/net v0.0.0-20190311183353-d8887717615a/go.mod h1:t9HGtf8HONx5eT2rtn7q6eTqICYqUVnKs3thJo3Qplg=
-golang.org/x/net v0.0.0-20190404232315-eb5bcb51f2a3/go.mod h1:t9HGtf8HONx5eT2rtn7q6eTqICYqUVnKs3thJo3Qplg=
-golang.org/x/net v0.0.0-20190620200207-3b0461eec859/go.mod h1:z5CRVTTTmAJ677TzLLGU+0bjPO0LkuOLi4/5GtJWs/s=
-golang.org/x/net v0.0.0-20190827160401-ba9fcec4b297/go.mod h1:z5CRVTTTmAJ677TzLLGU+0bjPO0LkuOLi4/5GtJWs/s=
-golang.org/x/net v0.0.0-20191007182048-72f939374954/go.mod h1:z5CRVTTTmAJ677TzLLGU+0bjPO0LkuOLi4/5GtJWs/s=
-golang.org/x/net v0.0.0-20200202094626-16171245cfb2/go.mod h1:z5CRVTTTmAJ677TzLLGU+0bjPO0LkuOLi4/5GtJWs/s=
-golang.org/x/net v0.0.0-20200226121028-0de0cce0169b/go.mod h1:z5CRVTTTmAJ677TzLLGU+0bjPO0LkuOLi4/5GtJWs/s=
-golang.org/x/net v0.0.0-20201010224723-4f7140c49acb/go.mod h1:sp8m0HH+o8qH0wwXwYZr8TS3Oi6o0r6Gce1SSxlDquU=
-golang.org/x/net v0.0.0-20201021035429-f5854403a974/go.mod h1:sp8m0HH+o8qH0wwXwYZr8TS3Oi6o0r6Gce1SSxlDquU=
-golang.org/x/net v0.0.0-20201110031124-69a78807bb2b/go.mod h1:sp8m0HH+o8qH0wwXwYZr8TS3Oi6o0r6Gce1SSxlDquU=
-golang.org/x/net v0.0.0-20201216054612-986b41b23924/go.mod h1:m0MpNAwzfU5UDzcl9v0D8zg8gWTRqZa9RBIspLL5mdg=
-golang.org/x/net v0.0.0-20201224014010-6772e930b67b/go.mod h1:m0MpNAwzfU5UDzcl9v0D8zg8gWTRqZa9RBIspLL5mdg=
-golang.org/x/net v0.0.0-20210119194325-5f4716e94777/go.mod h1:m0MpNAwzfU5UDzcl9v0D8zg8gWTRqZa9RBIspLL5mdg=
-golang.org/x/net v0.0.0-20210226172049-e18ecbb05110/go.mod h1:m0MpNAwzfU5UDzcl9v0D8zg8gWTRqZa9RBIspLL5mdg=
-golang.org/x/net v0.0.0-20210525063256-abc453219eb5/go.mod h1:9nx3DQGgdP8bBQD5qxJ1jj9UTztislL4KSBs9R2vV5Y=
-golang.org/x/net v0.0.0-20210805182204-aaa1db679c0d/go.mod h1:9nx3DQGgdP8bBQD5qxJ1jj9UTztislL4KSBs9R2vV5Y=
-golang.org/x/net v0.0.0-20210928044308-7d9f5e0b762b/go.mod h1:9nx3DQGgdP8bBQD5qxJ1jj9UTztislL4KSBs9R2vV5Y=
-golang.org/x/net v0.0.0-20211020060615-d418f374d309/go.mod h1:9nx3DQGgdP8bBQD5qxJ1jj9UTztislL4KSBs9R2vV5Y=
-golang.org/x/net v0.0.0-20211209124913-491a49abca63/go.mod h1:9nx3DQGgdP8bBQD5qxJ1jj9UTztislL4KSBs9R2vV5Y=
-golang.org/x/net v0.0.0-20220107192237-5cfca573fb4d/go.mod h1:9nx3DQGgdP8bBQD5qxJ1jj9UTztislL4KSBs9R2vV5Y=
-golang.org/x/net v0.41.0 h1:vBTly1HeNPEn3wtREYfy4GZ/NECgw2Cnl+nK6Nz3uvw=
-golang.org/x/net v0.41.0/go.mod h1:B/K4NNqkfmg07DQYrbwvSluqCJOOXwUjeb/5lOisjbA=
-golang.org/x/oauth2 v0.30.0 h1:dnDm7JmhM45NNpd8FDDeLhK6FwqbOf4MLCM9zb1BOHI=
-golang.org/x/oauth2 v0.30.0/go.mod h1:B++QgG3ZKulg6sRPGD/mqlHQs5rB3Ml9erfeDY7xKlU=
-golang.org/x/sync v0.0.0-20190423024810-112230192c58/go.mod h1:RxMgew5VJxzue5/jJTE5uejpjVlOe/izrB70Jof72aM=
-golang.org/x/sync v0.0.0-20190911185100-cd5d95a43a6e/go.mod h1:RxMgew5VJxzue5/jJTE5uejpjVlOe/izrB70Jof72aM=
-golang.org/x/sync v0.0.0-20201020160332-67f06af15bc9/go.mod h1:RxMgew5VJxzue5/jJTE5uejpjVlOe/izrB70Jof72aM=
+golang.org/x/mod v0.38.0 h1:MECBjubtXD7yj4HrhIUcywNaGeNVUdfVnxmPajOk4yk=
+golang.org/x/mod v0.38.0/go.mod h1:V6Xz0pq8TQ3dGqVQ1FVHuelZpAL0uNhSkk9ogYP3c40=
+golang.org/x/net v0.0.0-20190503192946-f4e77d36d62c/go.mod h1:t9HGtf8HONx5eT2rtn7q6eTqICYqUVnKs3thJo3Qplg=
+golang.org/x/net v0.0.0-20220923203811-8be639271d50/go.mod h1:YDH+HFinaLZZlnHAfSS6ZXJJ9M9t4Dl22yv3iI2vPwk=
+golang.org/x/net v0.57.0 h1:K5+3DljvIuDG9/Jv9rvyMywYNFCQ9RSUY6OOTTkT+tE=
+golang.org/x/net v0.57.0/go.mod h1:KpXc8iv+r3XplLAG/f7Jsf9RPszJzdR0f58q9vGOuEU=
+golang.org/x/oauth2 v0.36.0 h1:peZ/1z27fi9hUOFCAZaHyrpWG5lwe0RJEEEeH0ThlIs=
+golang.org/x/oauth2 v0.36.0/go.mod h1:YDBUJMTkDnJS+A4BP4eZBjCqtokkg1hODuPjwiGPO7Q=
 golang.org/x/sync v0.0.0-20210220032951-036812b2e83c/go.mod h1:RxMgew5VJxzue5/jJTE5uejpjVlOe/izrB70Jof72aM=
-golang.org/x/sync v0.15.0 h1:KWH3jNZsfyT6xfAfKiz6MRNmd46ByHDYaZ7KSkCtdW8=
-golang.org/x/sync v0.15.0/go.mod h1:1dzgHSNfp02xaA81J2MS99Qcpr2w7fw1gpm99rleRqA=
-golang.org/x/sys v0.0.0-20180905080454-ebe1bf3edb33/go.mod h1:STP8DvDyc/dI5b8T5hshtkjS+E42TnysNCUPdjciGhY=
+golang.org/x/sync v0.0.0-20220923202941-7f9b1623fab7/go.mod h1:RxMgew5VJxzue5/jJTE5uejpjVlOe/izrB70Jof72aM=
+golang.org/x/sync v0.22.0 h1:SZjpbeLmrCk4xhRSZFNZW5gFUeCeFgjekvI/+gfScek=
+golang.org/x/sync v0.22.0/go.mod h1:9xrNwdLfx4jkKbNva9FpL6vEN7evnE43NNNJQ2LF3+0=
 golang.org/x/sys v0.0.0-20190215142949-d0b11bdaac8a/go.mod h1:STP8DvDyc/dI5b8T5hshtkjS+E42TnysNCUPdjciGhY=
-golang.org/x/sys v0.0.0-20190312061237-fead79001313/go.mod h1:h1NjWce9XRLGQEsW7wpKNCjG9DtNlClVuFLEZdDNbEs=
-golang.org/x/sys v0.0.0-20190411185658-b44545bcd369/go.mod h1:h1NjWce9XRLGQEsW7wpKNCjG9DtNlClVuFLEZdDNbEs=
-golang.org/x/sys v0.0.0-20190412213103-97732733099d/go.mod h1:h1NjWce9XRLGQEsW7wpKNCjG9DtNlClVuFLEZdDNbEs=
-golang.org/x/sys v0.0.0-20190507160741-ecd444e8653b/go.mod h1:h1NjWce9XRLGQEsW7wpKNCjG9DtNlClVuFLEZdDNbEs=
-golang.org/x/sys v0.0.0-20190826190057-c7b8b68b1456/go.mod h1:h1NjWce9XRLGQEsW7wpKNCjG9DtNlClVuFLEZdDNbEs=
-golang.org/x/sys v0.0.0-20191008105621-543471e840be/go.mod h1:h1NjWce9XRLGQEsW7wpKNCjG9DtNlClVuFLEZdDNbEs=
-golang.org/x/sys v0.0.0-20200202164722-d101bd2416d5/go.mod h1:h1NjWce9XRLGQEsW7wpKNCjG9DtNlClVuFLEZdDNbEs=
-golang.org/x/sys v0.0.0-20200930185726-fdedc70b468f/go.mod h1:h1NjWce9XRLGQEsW7wpKNCjG9DtNlClVuFLEZdDNbEs=
-golang.org/x/sys v0.0.0-20201009025420-dfb3f7c4e634/go.mod h1:h1NjWce9XRLGQEsW7wpKNCjG9DtNlClVuFLEZdDNbEs=
-golang.org/x/sys v0.0.0-20201118182958-a01c418693c7/go.mod h1:h1NjWce9XRLGQEsW7wpKNCjG9DtNlClVuFLEZdDNbEs=
-golang.org/x/sys v0.0.0-20201119102817-f84b799fce68/go.mod h1:h1NjWce9XRLGQEsW7wpKNCjG9DtNlClVuFLEZdDNbEs=
-golang.org/x/sys v0.0.0-20201218084310-7d0127a74742/go.mod h1:h1NjWce9XRLGQEsW7wpKNCjG9DtNlClVuFLEZdDNbEs=
-golang.org/x/sys v0.0.0-20210110051926-789bb1bd4061/go.mod h1:h1NjWce9XRLGQEsW7wpKNCjG9DtNlClVuFLEZdDNbEs=
-golang.org/x/sys v0.0.0-20210119212857-b64e53b001e4/go.mod h1:h1NjWce9XRLGQEsW7wpKNCjG9DtNlClVuFLEZdDNbEs=
-golang.org/x/sys v0.0.0-20210123111255-9b0068b26619/go.mod h1:h1NjWce9XRLGQEsW7wpKNCjG9DtNlClVuFLEZdDNbEs=
-golang.org/x/sys v0.0.0-20210124154548-22da62e12c0c/go.mod h1:h1NjWce9XRLGQEsW7wpKNCjG9DtNlClVuFLEZdDNbEs=
-golang.org/x/sys v0.0.0-20210216163648-f7da38b97c65/go.mod h1:h1NjWce9XRLGQEsW7wpKNCjG9DtNlClVuFLEZdDNbEs=
-golang.org/x/sys v0.0.0-20210305230114-8fe3ee5dd75b/go.mod h1:h1NjWce9XRLGQEsW7wpKNCjG9DtNlClVuFLEZdDNbEs=
-golang.org/x/sys v0.0.0-20210423082822-04245dca01da/go.mod h1:h1NjWce9XRLGQEsW7wpKNCjG9DtNlClVuFLEZdDNbEs=
-golang.org/x/sys v0.0.0-20210525143221-35b2ab0089ea/go.mod h1:oPkhp1MJrh7nUepCBck5+mAzfO9JrbApNNgaTdGDITg=
-golang.org/x/sys v0.0.0-20210809222454-d867a43fc93e/go.mod h1:oPkhp1MJrh7nUepCBck5+mAzfO9JrbApNNgaTdGDITg=
-golang.org/x/sys v0.0.0-20210906170528-6f6e22806c34/go.mod h1:oPkhp1MJrh7nUepCBck5+mAzfO9JrbApNNgaTdGDITg=
-golang.org/x/sys v0.0.0-20210927094055-39ccf1dd6fa6/go.mod h1:oPkhp1MJrh7nUepCBck5+mAzfO9JrbApNNgaTdGDITg=
-golang.org/x/sys v0.0.0-20211025201205-69cdffdb9359/go.mod h1:oPkhp1MJrh7nUepCBck5+mAzfO9JrbApNNgaTdGDITg=
-golang.org/x/sys v0.0.0-20211210111614-af8b64212486/go.mod h1:oPkhp1MJrh7nUepCBck5+mAzfO9JrbApNNgaTdGDITg=
-golang.org/x/sys v0.0.0-20211216021012-1d35b9e2eb4e/go.mod h1:oPkhp1MJrh7nUepCBck5+mAzfO9JrbApNNgaTdGDITg=
-golang.org/x/sys v0.0.0-20220715151400-c0bba94af5f8/go.mod h1:oPkhp1MJrh7nUepCBck5+mAzfO9JrbApNNgaTdGDITg=
+golang.org/x/sys v0.0.0-20210615035016-665e8c7367d1/go.mod h1:oPkhp1MJrh7nUepCBck5+mAzfO9JrbApNNgaTdGDITg=
+golang.org/x/sys v0.0.0-20220319134239-a9b59b0215f8/go.mod h1:oPkhp1MJrh7nUepCBck5+mAzfO9JrbApNNgaTdGDITg=
+golang.org/x/sys v0.0.0-20220728004956-3c1f35247d10/go.mod h1:oPkhp1MJrh7nUepCBck5+mAzfO9JrbApNNgaTdGDITg=
 golang.org/x/sys v0.2.0/go.mod h1:oPkhp1MJrh7nUepCBck5+mAzfO9JrbApNNgaTdGDITg=
 golang.org/x/sys v0.6.0/go.mod h1:oPkhp1MJrh7nUepCBck5+mAzfO9JrbApNNgaTdGDITg=
 golang.org/x/sys v0.10.0/go.mod h1:oPkhp1MJrh7nUepCBck5+mAzfO9JrbApNNgaTdGDITg=
-golang.org/x/sys v0.33.0 h1:q3i8TbbEz+JRD9ywIRlyRAQbM0qF7hu24q3teo2hbuw=
-golang.org/x/sys v0.33.0/go.mod h1:BJP2sWEmIv4KK5OTEluFJCKSidICx8ciO85XgH3Ak8k=
-golang.org/x/term v0.0.0-20201126162022-7de9c90e9dd1/go.mod h1:bj7SfCRtBDWHUb9snDiAeCFNEtKQo2Wmx5Cou7ajbmo=
-golang.org/x/term v0.32.0 h1:DR4lr0TjUs3epypdhTOkMmuF5CDFJ/8pOnbzMZPQ7bg=
-golang.org/x/term v0.32.0/go.mod h1:uZG1FhGx848Sqfsq4/DlJr3xGGsYMu/L5GW4abiaEPQ=
+golang.org/x/sys v0.47.0 h1:o7XGOvZQCADBQQ4Y7VNq2dRWQR7JmOUW8Kxx4ZsNgWs=
+golang.org/x/sys v0.47.0/go.mod h1:4GL1E5IUh+htKOUEOaiffhrAeqysfVGipDYzABqnCmw=
+golang.org/x/term v0.0.0-20210927222741-03fcf44c2211/go.mod h1:jbD1KX2456YbFQfuXm/mYQcufACuNUgVhRMnK/tPxf8=
+golang.org/x/term v0.45.0 h1:NwWyBmoJCbfTHpxrWoZ9C6/VxOf7ic219I8xZZFdrf0=
+golang.org/x/term v0.45.0/go.mod h1:9aqxs0blBcrm/n0L9QW0aRVD+ktan8ssZromtqJC43w=
 golang.org/x/text v0.3.0/go.mod h1:NqM8EUOU14njkJ3fqMW+pc6Ldnwhi/IjpwHt7yyuwOQ=
-golang.org/x/text v0.3.3/go.mod h1:5Zoc/QRtKVWzQhOtBMvqHzDpF6irO9z98xDceosuGiQ=
-golang.org/x/text v0.3.6/go.mod h1:5Zoc/QRtKVWzQhOtBMvqHzDpF6irO9z98xDceosuGiQ=
-golang.org/x/text v0.26.0 h1:P42AVeLghgTYr4+xUnTRKDMqpar+PtX7KWuNQL21L8M=
-golang.org/x/text v0.26.0/go.mod h1:QK15LZJUUQVJxhz7wXgxSy/CJaTFjd0G+YLonydOVQA=
-golang.org/x/time v0.9.0 h1:EsRrnYcQiGH+5FfbgvV4AP7qEZstoyrHB0DzarOQ4ZY=
-golang.org/x/time v0.9.0/go.mod h1:3BpzKBy/shNhVucY/MWOyx10tF3SFh9QdLuxbVysPQM=
+golang.org/x/text v0.3.7/go.mod h1:u+2+/6zg+i71rQMx5EYifcz6MCKuco9NR6JIITiCfzQ=
+golang.org/x/text v0.40.0 h1:Ub2Z6/xjgF1WrYQz2nuITOEegKFtiIy+rieRJ5lHZKs=
+golang.org/x/text v0.40.0/go.mod h1:hpnzDAfGV753zIKo+wk3u1bVKCGPbrnF7+7LBF/UHVY=
+golang.org/x/time v0.14.0 h1:MRx4UaLrDotUKUdCIqzPC48t1Y9hANFKIRpNx+Te8PI=
+golang.org/x/time v0.14.0/go.mod h1:eL/Oa2bBBK0TkX57Fyni+NgnyQQN4LitPmob2Hjnqw4=
 golang.org/x/tools v0.0.0-20180917221912-90fa682c2a6e/go.mod h1:n7NCudcB/nEzxVGmLbDWY5pfWTLqBcC2KZ6jyYvM4mQ=
-golang.org/x/tools v0.0.0-20191119224855-298f0cb1881e/go.mod h1:b+2E5dAYhXwXZwtnZ6UAqBI28+e2cm9otk0dWdXHAEo=
-golang.org/x/tools v0.0.0-20200619180055-7c47624df98f/go.mod h1:EkVYQZoAsY45+roYkvgYkIh4xh/qjgUK9TdY2XT94GE=
-golang.org/x/tools v0.0.0-20210106214847-113979e3529a/go.mod h1:emZCQorbCU4vsT4fOWvOPXz4eW1wZW4PmDk9uLelYpA=
-golang.org/x/tools v0.1.0/go.mod h1:xkSsbof2nBLbhDlRMhhhyNLN/zl3eTqcnHD5viDpcZ0=
-golang.org/x/tools v0.1.7/go.mod h1:LGqMHiF4EqQNHR1JncWGqT5BVaXmza+X+BDGol+dOxo=
-golang.org/x/tools v0.33.0 h1:4qz2S3zmRxbGIhDIAgjxvFutSvH5EfnsYrRBj0UI0bc=
-golang.org/x/tools v0.33.0/go.mod h1:CIJMaWEY88juyUfo7UbgPqbC8rU2OqfAV1h2Qp0oMYI=
-golang.org/x/xerrors v0.0.0-20190717185122-a985d3407aa7/go.mod h1:I/5z698sn9Ka8TeJc9MKroUUfqBBauWjQqLJ2OPfmY0=
-golang.org/x/xerrors v0.0.0-20191011141410-1b5146add898/go.mod h1:I/5z698sn9Ka8TeJc9MKroUUfqBBauWjQqLJ2OPfmY0=
+golang.org/x/tools v0.48.0 h1:3+hClM1aLL5mjMKm5ovokw9epgRXPuu2tILgismM6RE=
+golang.org/x/tools v0.48.0/go.mod h1:08xX0orndb/F7jJxGDicx061tyd5pcMto75YMAXr6lk=
 golang.org/x/xerrors v0.0.0-20191204190536-9bdfabe68543/go.mod h1:I/5z698sn9Ka8TeJc9MKroUUfqBBauWjQqLJ2OPfmY0=
-golang.org/x/xerrors v0.0.0-20200804184101-5ec99f83aff1/go.mod h1:I/5z698sn9Ka8TeJc9MKroUUfqBBauWjQqLJ2OPfmY0=
 golang.zx2c4.com/wireguard v0.0.0-20231211153847-12269c276173 h1:/jFs0duh4rdb8uIfPMv78iAJGcPKDeqAFnaLBropIC4=
 golang.zx2c4.com/wireguard v0.0.0-20231211153847-12269c276173/go.mod h1:tkCQ4FQXmpAgYVh++1cq16/dH4QJtmvpRv19DWGAHSA=
 golang.zx2c4.com/wireguard/wgctrl v0.0.0-20241231184526-a9ab2273dd10 h1:3GDAcqdIg1ozBNLgPy4SLT84nfcBjr6rhGtXYtrkWLU=
 golang.zx2c4.com/wireguard/wgctrl v0.0.0-20241231184526-a9ab2273dd10/go.mod h1:T97yPqesLiNrOYxkwmhMI0ZIlJDm+p0PMR8eRVeR5tQ=
-google.golang.org/genproto/googleapis/api v0.0.0-20250603155806-513f23925822 h1:oWVWY3NzT7KJppx2UKhKmzPq4SRe0LdCijVRwvGeikY=
-google.golang.org/genproto/googleapis/api v0.0.0-20250603155806-513f23925822/go.mod h1:h3c4v36UTKzUiuaOKQ6gr3S+0hovBtUrXzTG/i3+XEc=
-google.golang.org/genproto/googleapis/rpc v0.0.0-20250603155806-513f23925822 h1:fc6jSaCT0vBduLYZHYrBBNY4dsWuvgyff9noRNDdBeE=
-google.golang.org/genproto/googleapis/rpc v0.0.0-20250603155806-513f23925822/go.mod h1:qQ0YXyHHx3XkvlzUtpXDkS29lDSafHMZBAZDc03LQ3A=
-google.golang.org/grpc v1.73.0 h1:VIWSmpI2MegBtTuFt5/JWy2oXxtjJ/e89Z70ImfD2ok=
-google.golang.org/grpc v1.73.0/go.mod h1:50sbHOUqWoCQGI8V2HQLJM0B+LMlIUjNSZmow7EVBQc=
-google.golang.org/protobuf v1.36.6 h1:z1NpPI8ku2WgiWnf+t9wTPsn6eP1L7ksHUlkfLvd9xY=
-google.golang.org/protobuf v1.36.6/go.mod h1:jduwjTPXsFjZGTmRluh+L6NjiWu7pchiJ2/5YcXBHnY=
+gonum.org/v1/gonum v0.17.0 h1:VbpOemQlsSMrYmn7T2OUvQ4dqxQXU+ouZFQsZOx50z4=
+gonum.org/v1/gonum v0.17.0/go.mod h1:El3tOrEuMpv2UdMrbNlKEh9vd86bmQ6vqIcDwxEOc1E=
+google.golang.org/genproto/googleapis/api v0.0.0-20260526163538-3dc84a4a5aaa h1:Kjn0N0tCrDgiAFW+lGO4JZ3ck44CehvJQMAwj9QF0G8=
+google.golang.org/genproto/googleapis/api v0.0.0-20260526163538-3dc84a4a5aaa/go.mod h1:q4lMZS6kskjT5HvCPrnnypcDPVJqT/f4nfxmkE7gryY=
+google.golang.org/genproto/googleapis/rpc v0.0.0-20260526163538-3dc84a4a5aaa h1:mZHHdPZl0dbGHCflZgAq/Q468DWVFcU2whhB2KAo8fk=
+google.golang.org/genproto/googleapis/rpc v0.0.0-20260526163538-3dc84a4a5aaa/go.mod h1:4Hqkh8ycfw05ld/3BWL7rJOSfebL2Q+DVDeRgYgxUU8=
+google.golang.org/grpc v1.83.0 h1:JeNZEKJFbQxArAMl+hiytHauacDNqJUllNfmIMmpqnQ=
+google.golang.org/grpc v1.83.0/go.mod h1:kDyl6SKsiHKt0uylY5gtn5cEjkrIOhQOGDgIc4JGwzQ=
+google.golang.org/protobuf v1.36.12-0.20260120151049-f2248ac996af h1:+5/Sw3GsDNlEmu7TfklWKPdQ0Ykja5VEmq2i817+jbI=
+google.golang.org/protobuf v1.36.12-0.20260120151049-f2248ac996af/go.mod h1:HTf+CrKn2C3g5S8VImy6tdcUvCska2kB7j23XfzDpco=
 gopkg.in/check.v1 v0.0.0-20161208181325-20d25e280405/go.mod h1:Co6ibVJAznAaIkqp8huTwlJQCZ016jof/cbN4VW5Yz0=
 gopkg.in/check.v1 v1.0.0-20201130134442-10cb98267c6c h1:Hei/4ADfdWqJk1ZMxUNpqntNwaWcugrBjAiHlqqRiVk=
 gopkg.in/check.v1 v1.0.0-20201130134442-10cb98267c6c/go.mod h1:JHkPIbrfpd72SG/EVd6muEfDQjcINNoR0C8j2r3qZ4Q=
-gopkg.in/evanphx/json-patch.v4 v4.12.0 h1:n6jtcsulIzXPJaxegRbvFNNrZDjbij7ny3gmSPG+6V4=
-gopkg.in/evanphx/json-patch.v4 v4.12.0/go.mod h1:p8EYWUEYMpynmqDbY58zCKCFZw8pRWMG4EsWvDvM72M=
+gopkg.in/evanphx/json-patch.v4 v4.13.0 h1:czT3CmqEaQ1aanPc5SdlgQrrEIb8w/wwCvWWnfEbYzo=
+gopkg.in/evanphx/json-patch.v4 v4.13.0/go.mod h1:p8EYWUEYMpynmqDbY58zCKCFZw8pRWMG4EsWvDvM72M=
 gopkg.in/inf.v0 v0.9.1 h1:73M5CoZyi3ZLMOyDlQh031Cx6N9NDJ2Vvfl76EDAgDc=
 gopkg.in/inf.v0 v0.9.1/go.mod h1:cWUDdTG/fYaXco+Dcufb5Vnc6Gp2YChqWtbxRZE0mXw=
-gopkg.in/ini.v1 v1.67.0 h1:Dgnx+6+nfE+IfzjUEISNeydPJh9AXNNsWbGP9KzCsOA=
-gopkg.in/ini.v1 v1.67.0/go.mod h1:pNLf8WUiyNEtQjuu5G5vTm06TEv9tsIgeAvK8hOrP4k=
-gopkg.in/yaml.v2 v2.4.0 h1:D8xgwECY7CYvx+Y2n4sBz93Jn9JRvxdiyyo8CTfuKaY=
-gopkg.in/yaml.v2 v2.4.0/go.mod h1:RDklbk79AGWmwhnvt/jBztapEOGDOx6ZbXqjP6csGnQ=
-gopkg.in/yaml.v3 v3.0.0-20200313102051-9f266ea9e77c/go.mod h1:K4uyk7z7BCEPqu6E+C64Yfv1cQ7kz7rIZviUmN+EgEM=
 gopkg.in/yaml.v3 v3.0.1 h1:fxVm/GzAzEWqLHuvctI91KS9hhNmmWOoWu0XTYJS7CA=
 gopkg.in/yaml.v3 v3.0.1/go.mod h1:K4uyk7z7BCEPqu6E+C64Yfv1cQ7kz7rIZviUmN+EgEM=
 gotest.tools/v3 v3.4.0 h1:ZazjZUfuVeZGLAmlKKuyv3IKP5orXcwtOwDQH6YVr6o=
 gotest.tools/v3 v3.4.0/go.mod h1:CtbdzLSsqVhDgMtKsx03ird5YTGB3ar27v0u/yKBW5g=
-honnef.co/go/tools v0.2.1/go.mod h1:lPVVZ2BS5TfnjLyizF7o7hv7j9/L+8cZY2hLyjP9cGY=
-honnef.co/go/tools v0.2.2/go.mod h1:lPVVZ2BS5TfnjLyizF7o7hv7j9/L+8cZY2hLyjP9cGY=
-k8s.io/api v0.33.3 h1:SRd5t//hhkI1buzxb288fy2xvjubstenEKL9K51KBI8=
-k8s.io/api v0.33.3/go.mod h1:01Y/iLUjNBM3TAvypct7DIj0M0NIZc+PzAHCIo0CYGE=
-k8s.io/apimachinery v0.33.3 h1:4ZSrmNa0c/ZpZJhAgRdcsFcZOw1PQU1bALVQ0B3I5LA=
-k8s.io/apimachinery v0.33.3/go.mod h1:BHW0YOu7n22fFv/JkYOEfkUYNRN0fj0BlvMFWA7b+SM=
-k8s.io/client-go v0.33.3 h1:M5AfDnKfYmVJif92ngN532gFqakcGi6RvaOF16efrpA=
-k8s.io/client-go v0.33.3/go.mod h1:luqKBQggEf3shbxHY4uVENAxrDISLOarxpTKMiUuujg=
-k8s.io/klog/v2 v2.130.1 h1:n9Xl7H1Xvksem4KFG4PYbdQCQxqc/tTUyrgXaOhHSzk=
-k8s.io/klog/v2 v2.130.1/go.mod h1:3Jpz1GvMt720eyJH1ckRHK1EDfpxISzJ7I9OYgaDtPE=
-k8s.io/kube-openapi v0.0.0-20250318190949-c8a335a9a2ff h1:/usPimJzUKKu+m+TE36gUyGcf03XZEP0ZIKgKj35LS4=
-k8s.io/kube-openapi v0.0.0-20250318190949-c8a335a9a2ff/go.mod h1:5jIi+8yX4RIb8wk3XwBo5Pq2ccx4FP10ohkbSKCZoK8=
-k8s.io/utils v0.0.0-20241210054802-24370beab758 h1:sdbE21q2nlQtFh65saZY+rRM6x6aJJI8IUa1AmH/qa0=
-k8s.io/utils v0.0.0-20241210054802-24370beab758/go.mod h1:OLgZIPagt7ERELqWJFomSt595RzquPNLL48iOWgYOg0=
-modernc.org/cc/v4 v4.24.4 h1:TFkx1s6dCkQpd6dKurBNmpo+G8Zl4Sq/ztJ+2+DEsh0=
-modernc.org/cc/v4 v4.24.4/go.mod h1:uVtb5OGqUKpoLWhqwNQo/8LwvoiEBLvZXIQ/SmO6mL0=
-modernc.org/ccorpus2 v1.5.2 h1:Ui+4tc58mf/W+2arcYCJR903y3zl3ecsI7Fpaaqozyw=
-modernc.org/ccorpus2 v1.5.2/go.mod h1:Wifvo4Q/qS/h1aRoC2TffcHsnxwTikmi1AuLANuucJQ=
-modernc.org/mathutil v1.7.1 h1:GCZVGXdaN8gTqB1Mf/usp1Y/hSqgI2vAGGP4jZMCxOU=
-modernc.org/mathutil v1.7.1/go.mod h1:4p5IwJITfppl0G4sUEDtCr4DthTaT47/N3aT6MhfgJg=
-modernc.org/opt v0.1.4 h1:2kNGMRiUjrp4LcaPuLY2PzUfqM/w9N23quVwhKt5Qm8=
-modernc.org/opt v0.1.4/go.mod h1:03fq9lsNfvkYSfxrfUhZCWPk1lm4cq4N+Bh//bEtgns=
-modernc.org/sortutil v1.2.1 h1:+xyoGf15mM3NMlPDnFqrteY07klSFxLElE2PVuWIJ7w=
-modernc.org/sortutil v1.2.1/go.mod h1:7ZI3a3REbai7gzCLcotuw9AC4VZVpYMjDzETGsSMqJE=
-modernc.org/strutil v1.2.1 h1:UneZBkQA+DX2Rp35KcM69cSsNES9ly8mQWD71HKlOA0=
-modernc.org/strutil v1.2.1/go.mod h1:EHkiggD70koQxjVdSBM3JKM7k6L0FbGE5eymy9i3B9A=
-modernc.org/token v1.1.0 h1:Xl7Ap9dKaEs5kLoOQeQmPWevfnk/DM5qcLcYlA8ys6Y=
-modernc.org/token v1.1.0/go.mod h1:UGzOrNV1mAFSEB63lOFHIpNRUVMvYTc6yu1SMY/XTDM=
+k8s.io/api v0.36.3 h1:NxB+05W2UGqXWFXcLO0RB5cnqnUPP5v5sVlaOH0Iz4w=
+k8s.io/api v0.36.3/go.mod h1:JzLQKqRHC5+I8RVj/lS3lCg0mg6nWI9Fo/Sk3ElxHzg=
+k8s.io/apimachinery v0.36.3 h1:PkzMRBRG8joFD8EhCuQAtNPvJlxb82FwplP26HIzvAM=
+k8s.io/apimachinery v0.36.3/go.mod h1:cTSjBWgPe/6CQyBKzY/hDIRWCQQQeK0mfLbml0UYFHE=
+k8s.io/client-go v0.36.3 h1:M4JdVzXxYcZk4fGpfDdYnxSwhLKWCFoQsHW6t+z8Hfg=
+k8s.io/client-go v0.36.3/go.mod h1:gcPwr0c87vjjG6HB6pWEqOeuYVoXSsREjzux2j6GF30=
+k8s.io/klog/v2 v2.140.0 h1:Tf+J3AH7xnUzZyVVXhTgGhEKnFqye14aadWv7bzXdzc=
+k8s.io/klog/v2 v2.140.0/go.mod h1:o+/RWfJ6PwpnFn7OyAG3QnO47BFsymfEfrz6XyYSSp0=
+k8s.io/kube-openapi v0.0.0-20260317180543-43fb72c5454a h1:xCeOEAOoGYl2jnJoHkC3hkbPJgdATINPMAxaynU2Ovg=
+k8s.io/kube-openapi v0.0.0-20260317180543-43fb72c5454a/go.mod h1:uGBT7iTA6c6MvqUvSXIaYZo9ukscABYi2btjhvgKGZ0=
+k8s.io/utils v0.0.0-20260210185600-b8788abfbbc2 h1:AZYQSJemyQB5eRxqcPky+/7EdBj0xi3g0ZcxxJ7vbWU=
+k8s.io/utils v0.0.0-20260210185600-b8788abfbbc2/go.mod h1:xDxuJ0whA3d0I4mf/C4ppKHxXynQ+fxnkmQH0vTHnuk=
 pgregory.net/rapid v1.1.0 h1:CMa0sjHSru3puNx+J0MIAuiiEV4N0qj8/cMWGBBCsjw=
 pgregory.net/rapid v1.1.0/go.mod h1:PY5XlDGj0+V1FCq0o192FdRhpKHGTRIWBgqjDBTrq04=
-sigs.k8s.io/json v0.0.0-20241014173422-cfa47c3a1cc8 h1:gBQPwqORJ8d8/YNZWEjoZs7npUVDpVXUUOFfW6CgAqE=
-sigs.k8s.io/json v0.0.0-20241014173422-cfa47c3a1cc8/go.mod h1:mdzfpAEoE6DHQEN0uh9ZbOCuHbLK5wOm7dK4ctXE9Tg=
-sigs.k8s.io/kind v0.29.0 h1:3TpCsyh908IkXXpcSnsMjWdwdWjIl7o9IMZImZCWFnI=
-sigs.k8s.io/kind v0.29.0/go.mod h1:ldWQisw2NYyM6k64o/tkZng/1qQW7OlzcN5a8geJX3o=
-sigs.k8s.io/randfill v0.0.0-20250304075658-069ef1bbf016/go.mod h1:XeLlZ/jmk4i1HRopwe7/aU3H5n1zNUcX6TM94b3QxOY=
+sigs.k8s.io/json v0.0.0-20250730193827-2d320260d730 h1:IpInykpT6ceI+QxKBbEflcR5EXP7sU1kvOlxwZh5txg=
+sigs.k8s.io/json v0.0.0-20250730193827-2d320260d730/go.mod h1:mdzfpAEoE6DHQEN0uh9ZbOCuHbLK5wOm7dK4ctXE9Tg=
+sigs.k8s.io/kind v0.32.0 h1:p9hscbj98u/qyrjVpjId86LI70nQmbSsipV7wCG10Xk=
+sigs.k8s.io/kind v0.32.0/go.mod h1:FSqriGaoTPruiXWfRnUXNykF8r2t+fHtK0P0m1AbGF8=
 sigs.k8s.io/randfill v1.0.0 h1:JfjMILfT8A6RbawdsK2JXGBR5AQVfd+9TbzrlneTyrU=
 sigs.k8s.io/randfill v1.0.0/go.mod h1:XeLlZ/jmk4i1HRopwe7/aU3H5n1zNUcX6TM94b3QxOY=
-sigs.k8s.io/structured-merge-diff/v4 v4.6.0 h1:IUA9nvMmnKWcj5jl84xn+T5MnlZKThmUW1TdblaLVAc=
-sigs.k8s.io/structured-merge-diff/v4 v4.6.0/go.mod h1:dDy58f92j70zLsuZVuUX5Wp9vtxXpaZnkPGWeqDfCps=
-sigs.k8s.io/yaml v1.4.0 h1:Mk1wCc2gy/F0THH0TAp1QYyJNzRm2KCLy3o5ASXVI5E=
-sigs.k8s.io/yaml v1.4.0/go.mod h1:Ejl7/uTz7PSA4eKMyQCUTnhZYNmLIl+5c2lQPGR2BPY=
+sigs.k8s.io/structured-merge-diff/v6 v6.3.3 h1:u08YRbVUi59ri4YD6cg0UqNM4Dimn0sIl+wldcx5PYw=
+sigs.k8s.io/structured-merge-diff/v6 v6.3.3/go.mod h1:M3W8sfWvn2HhQDIbGWj3S099YozAsymCo/wrT5ohRUE=
+sigs.k8s.io/yaml v1.6.0 h1:G8fkbMSAFqgEFgh4b1wmtzDnioxFCUgTZhlbj5P9QYs=
+sigs.k8s.io/yaml v1.6.0/go.mod h1:796bPqUfzR/0jLAl6XjHl3Ck7MiyVv8dbTdyT3/pMf4=

main.go

@@ -1,6 +1,8 @@
 package main
 
 import (
+	"os"
+
 	"github.com/kube-vip/kube-vip/cmd"
 )
 
@@ -14,5 +16,5 @@ func main() {
 
 	cmd.Release.Version = Version
 	cmd.Release.Build = Build
-	cmd.Execute()
+	os.Exit(cmd.Execute())
 }

pkg/arp/arp.go

@@ -2,16 +2,19 @@ package arp
 
 import (
 	"context"
+	"fmt"
 	log "log/slog"
 	"sync"
 	"time"
 
 	"github.com/kube-vip/kube-vip/pkg/kubevip"
+	"github.com/kube-vip/kube-vip/pkg/utils"
 	"github.com/kube-vip/kube-vip/pkg/vip"
+	"github.com/vishvananda/netlink"
 )
 
 type Manager struct {
-	instances map[string]*Instance
+	instances sync.Map
 	config    *kubevip.Config
 }
 
@@ -23,9 +26,12 @@ type Instance struct {
 }
 
 func NewManager(config *kubevip.Config) *Manager {
+	if config.ArpBroadcastRate < 500 {
+		log.Warn("[ARP manager] arp broadcast rate is too low", "rate (ms)", config.ArpBroadcastRate, "setting to (ms)", "3000")
+		config.ArpBroadcastRate = 3000
+	}
 	return &Manager{
-		instances: make(map[string]*Instance),
-		config:    config,
+		config: config,
 	}
 }
 
@@ -42,10 +48,14 @@ func (i *Instance) Name() string {
 }
 
 func (m *Manager) Insert(instance *Instance) {
-	i, ok := m.instances[instance.Name()]
-	if !ok {
-		log.Info("inserting ARP/NDP instance", "name", instance.Name())
-		m.instances[instance.Name()] = instance
+	i, err := m.get(instance.Name())
+	if err != nil {
+		log.Error("[ARP manager] unable to insert instance", "err", err)
+		return
+	}
+	if i == nil {
+		log.Info("[ARP manager] inserting ARP/NDP instance", "name", instance.Name())
+		m.instances.Store(instance.Name(), instance)
 	} else {
 		i.mu.Lock()
 		defer i.mu.Unlock()
@@ -54,20 +64,49 @@ func (m *Manager) Insert(instance *Instance) {
 }
 
 func (m *Manager) Remove(instance *Instance) {
-	if i, ok := m.instances[instance.Name()]; ok {
+	m.RemoveWithIPDelete(instance, true)
+}
+
+// RemoveOnLeadershipLoss removes an ARP instance when leadership is lost
+func (m *Manager) RemoveOnLeadershipLoss(instance *Instance) {
+	// Use the inverse of PreserveVIPOnLeadershipLoss to decide whether to delete the IP
+	// If preserve is true, don't delete IP (deleteIP = false)
+	// If preserve is false, delete IP (deleteIP = true), This is the legacy behavior
+	deleteIP := !m.config.PreserveVIPOnLeadershipLoss
+	m.RemoveWithIPDelete(instance, deleteIP)
+}
+
+func (m *Manager) RemoveWithIPDelete(instance *Instance, deleteIP bool) {
+	i, err := m.get(instance.Name())
+	if err != nil {
+		log.Error("[ARP manager] unable to remove the instance", "err", err)
+		return
+	}
+	if i != nil {
 		i.mu.Lock()
 		defer i.mu.Unlock()
-		if i.counter > 1 {
-			i.counter--
-		} else {
-			log.Info("removing ARP/NDP instance", "name", instance.Name())
-			delete(m.instances, instance.Name())
+		i.counter--
+		if i.counter == 0 {
+			log.Info("[ARP manager] removing ARP/NDP instance", "name", instance.Name())
+			if deleteIP {
+				if _, err := instance.network.DeleteIP(); err != nil {
+					log.Error("failed to delete IP", "address", instance.network.IP(), "err", err)
+				}
+			}
+			m.instances.Delete(instance.Name())
 		}
+	} else {
+		log.Warn("[ARP manager] unable to remove the instance - instance not found", "name", instance.Name())
 	}
 }
 
 func (m *Manager) Count(name string) int {
-	if i, ok := m.instances[name]; ok {
+	i, err := m.get(name)
+	if err != nil {
+		log.Error("[ARP manager] unable to count instance", "err", err)
+		return -1
+	}
+	if i != nil {
 		i.mu.Lock()
 		defer i.mu.Unlock()
 		return i.counter
@@ -75,27 +114,81 @@ func (m *Manager) Count(name string) int {
 	return 0
 }
 
-func (m *Manager) StartAdvertisement(ctx context.Context) {
-	log.Info("Starting ARP/NDP advertisement")
+func (m *Manager) StartAdvertisement(ctx context.Context, killFunc func()) {
+	if m.config.LoseLeadership {
+		var wg sync.WaitGroup
+		defer wg.Wait()
+
+		log.Info("[ARP manager] starting watching network device", "interface", m.config.Interface)
+
+		duration := time.Duration(m.config.LoseLeadershipTimeoutSeconds) * time.Second
+		timeout := time.NewTimer(duration)
+		timeout.Stop()
+
+		wg.Go(func() {
+			select {
+			case <-timeout.C:
+				killFunc()
+			case <-ctx.Done():
+				return
+			}
+		})
+
+		wg.Go(func() {
+			if err := watch(ctx, m.config.Interface, func(s netlink.LinkOperState) {
+				if isUp(s) {
+					timeout.Stop()
+					return
+				}
+				timeout.Reset(duration)
+			}); err != nil {
+				log.Error("[ARP manager] stopped watching interface", "err", err)
+				killFunc()
+			}
+		})
+	}
+
+	log.Info("[ARP manager] starting ARP/NDP advertisement")
+
+	ticker := time.NewTicker(time.Duration(m.config.ArpBroadcastRate) * time.Millisecond)
+	defer ticker.Stop()
+
 	for {
 		select {
 		case <-ctx.Done(): // if cancel() execute
 			return
-		default:
-			for _, instance := range m.instances {
-				if instance.counter > 0 {
-					ensureIPAndSendGratuitous(instance)
+		case <-ticker.C: // send gratuitous ARP/NDP on each tick
+			m.instances.Range(func(_ any, instance any) bool {
+				if i, ok := instance.(*Instance); ok {
+					i.mu.Lock()
+					defer i.mu.Unlock()
+					if i.counter > 0 {
+						ensureIPAndSendGratuitous(i)
+					} else {
+						// this instance should not be advertised - delete the IP just in case...
+						if _, err := i.network.DeleteIP(); err != nil {
+							log.Error("[ARP manager] failed to delete IP", "address", i.network.IP(), "err", err)
+						}
+					}
 				}
-			}
+				return true
+			})
 		}
-		if m.config.ArpBroadcastRate < 500 {
-			log.Error("arp broadcast rate is too low", "rate (ms)", m.config.ArpBroadcastRate, "setting to (ms)", "3000")
-			m.config.ArpBroadcastRate = 3000
-		}
-		time.Sleep(time.Duration(m.config.ArpBroadcastRate) * time.Millisecond)
 	}
 }
 
+func (m *Manager) get(name string) (*Instance, error) {
+	i, exists := m.instances.Load(name)
+	if !exists {
+		return nil, nil
+	}
+	inst, ok := i.(*Instance)
+	if !ok {
+		return nil, fmt.Errorf("value for name %q is not of Instance pointer type", name)
+	}
+	return inst, nil
+}
+
 // ensureIPAndSendGratuitous - adds IP to the interface if missing, and send
 // either a gratuitous ARP or gratuitous NDP. Re-adds the interface if it is IPv6
 // and in a dadfailed state.
@@ -111,23 +204,27 @@ func ensureIPAndSendGratuitous(instance *Instance) {
 			log.Warn(err.Error())
 		}
 		if deleted {
-			log.Info("deleted and recreating address", "IP", ipString, "interface", iface)
-			// if _, err := instance.network.AddIP(false); err != nil {
-			// 	log.Error("failed to recreate address", "IP", ipString, "interface", iface)
-			// }
+			log.Info("deleted and recreating address with NODAD flag to skip DAD", "IP", ipString, "interface", iface)
+			// Re-add immediately without DAD check since we're recovering from DADFAILED
+			// The AddIP function will set IFA_F_NODAD flag for IPv6 addresses when skipDAD=true
+			if _, err := instance.network.AddIP(false, true); err != nil {
+				log.Error("failed to recreate address after DADFAILED", "IP", ipString, "interface", iface, "err", err)
+			} else {
+				log.Info("successfully recreated address after DADFAILED recovery", "IP", ipString, "interface", iface)
+			}
 		}
+		// Return early after DADFAILED recovery to avoid double IP addition
+		return
 	}
 
-	// Ensure the address exists on the interface before attempting to ARP
-	// if instance.network.HasEndpoints() {
-	if added, err := instance.network.AddIP(true); err != nil {
+	// Normal case: add IP with precheck and normal DAD process
+	if added, err := instance.network.AddIP(true, false); err != nil {
 		log.Warn(err.Error())
 	} else if added {
 		log.Warn("Re-applied the VIP configuration", "ip", ipString, "interface", iface)
 	}
-	// }
 
-	if vip.IsIPv6(ipString) {
+	if utils.IsIPv6(ipString) {
 		// Gratuitous NDP, will broadcast new MAC <-> IPv6 address
 		if instance.ndp == nil {
 			log.Error("NDP responder was not created")
@@ -137,7 +234,6 @@ func ensureIPAndSendGratuitous(instance *Instance) {
 				log.Warn(err.Error())
 			}
 		}
-
 	} else {
 		// Gratuitous ARP, will broadcast to new MAC <-> IPv4 address
 		err := vip.ARPSendGratuitous(ipString, iface)
@@ -145,5 +241,52 @@ func ensureIPAndSendGratuitous(instance *Instance) {
 			log.Warn(err.Error())
 		}
 	}
-
+}
+
+// watch subscribing to the network interface events and calls handler
+func watch(ctx context.Context, interfaceName string, operStateHandler func(netlink.LinkOperState)) error {
+	ifname, err := netlink.LinkByName(interfaceName)
+	if err != nil {
+		return fmt.Errorf("failed to watch interface %q: %w", interfaceName, err)
+	}
+
+	// verify if this interface is physical device
+	if _, ok := ifname.(*netlink.Device); !ok {
+		return fmt.Errorf("interface %s is not physical, ignoring", interfaceName)
+	}
+
+	events := make(chan netlink.LinkUpdate)
+	done := make(chan struct{})
+
+	if err := netlink.LinkSubscribe(events, done); err != nil {
+		return fmt.Errorf("failed to subscribe to the interface events: %w", err)
+	}
+	defer close(done)
+
+	//  handle initial state
+	operStateHandler(ifname.Attrs().OperState)
+
+	for {
+		select {
+		case <-ctx.Done():
+			return ctx.Err()
+		case event, ok := <-events:
+			if !ok {
+				return fmt.Errorf("interface events channel closed")
+			}
+
+			attrs := event.Attrs()
+			// LinkSubscribe captures events for all network devices found
+			// so we only care about vip interface
+			if ifname.Attrs().Name != attrs.Name {
+				continue
+			}
+			log.Debug("handling device change", "state", attrs.OperState)
+			operStateHandler(attrs.OperState)
+		}
+	}
+}
+
+func isUp(operState netlink.LinkOperState) bool {
+	return operState == netlink.OperUp
 }

pkg/backend/backend.go

@@ -1,14 +1,15 @@
 package backend
 
 import (
+	"context"
 	"fmt"
+	"sync"
 	"time"
 
 	log "log/slog"
 
 	"github.com/kube-vip/kube-vip/pkg/k8s"
 	"github.com/kube-vip/kube-vip/pkg/utils"
-	"github.com/kube-vip/kube-vip/pkg/vip"
 	"k8s.io/client-go/kubernetes"
 	"k8s.io/client-go/rest"
 )
@@ -21,6 +22,21 @@ type Entry struct {
 
 type Map map[Entry]bool
 
+// kubeConfigPath is an explicitly configured kubeconfig used by Check when
+// set; static pod deployments configure it since neither admin.conf nor
+// in-cluster config are available there.
+var (
+	kubeConfigPath string
+	pathMtx        sync.Mutex
+)
+
+// SetKubeConfigPath configures the kubeconfig used by backend health checks.
+func SetKubeConfigPath(path string) {
+	pathMtx.Lock()
+	defer pathMtx.Unlock()
+	kubeConfigPath = path
+}
+
 func (e *Entry) Check() bool {
 	var client *kubernetes.Clientset
 	var err error
@@ -31,13 +47,19 @@ func (e *Entry) Check() bool {
 	// homeConfigPath := filepath.Join(os.Getenv("HOME"), ".kube", "config")
 
 	var k8sAddr string
-	if vip.IsIPv4(e.Addr) {
-		k8sAddr = fmt.Sprintf("%s:%v", e.Addr, e.Port)
-	} else {
+	if utils.IsIPv6(e.Addr) {
 		k8sAddr = fmt.Sprintf("[%s]:%v", e.Addr, e.Port)
+	} else {
+		k8sAddr = fmt.Sprintf("%s:%v", e.Addr, e.Port)
 	}
 
 	switch {
+	case kubeConfigPath != "" && utils.FileExists(kubeConfigPath):
+		config, err = k8s.NewRestConfig(kubeConfigPath, false, k8sAddr)
+		if err != nil {
+			log.Error("create k8s REST config", "path", kubeConfigPath, "err", err)
+			return false
+		}
 	case utils.FileExists(adminConfigPath):
 		config, err = k8s.NewRestConfig(adminConfigPath, false, k8sAddr)
 		if err != nil {
@@ -66,7 +88,7 @@ func (e *Entry) Check() bool {
 	return true
 }
 
-func Watch(tickAction func(), interval int, stop chan struct{}) {
+func Watch(ctx context.Context, interval int, tickAction func()) {
 	if interval <= 0 {
 		interval = 5
 	}
@@ -76,13 +98,10 @@ func Watch(tickAction func(), interval int, stop chan struct{}) {
 
 	for {
 		select {
-		case <-stop:
-			ticker.Stop()
+		case <-ctx.Done():
 			return
 		case <-ticker.C:
-			ticker.Stop()
 			tickAction()
-			ticker.Reset(time.Second * time.Duration(interval))
 		}
 	}
 }

pkg/bgp/hosts.go

@@ -3,46 +3,80 @@ package bgp
 import (
 	"context"
 	"fmt"
+	log "log/slog"
 	"net"
 
-	api "github.com/osrg/gobgp/v3/api"
+	"github.com/osrg/gobgp/v4/pkg/apiutil"
 )
 
 // AddHost will update peers of a host
-func (b *Server) AddHost(addr string) (err error) {
-	ip, _, err := net.ParseCIDR(addr)
-	if err != nil {
-		return err
+func (b *Server) AddHost(ctx context.Context, addr string, object string) error {
+	b.mtx.Lock()
+	defer b.mtx.Unlock()
+
+	objects, exists := b.tracker[addr]
+
+	if !exists {
+		b.tracker[addr] = make(map[string]bool)
+		objects = b.tracker[addr]
+
+		ip, _, err := net.ParseCIDR(addr)
+		if err != nil {
+			return err
+		}
+
+		p := b.getPath(ip)
+		if p == nil {
+			return fmt.Errorf("failed to get path for %v", ip)
+		}
+
+		if _, err := b.s.AddPath(apiutil.AddPathRequest{
+			Paths: []*apiutil.Path{p},
+		}); err != nil {
+			return err
+		}
+		log.Debug("[BGP] added host", "addr", addr, "cnt", len(objects)+1, "object", object)
 	}
 
-	p := b.getPath(ip)
-	if p == nil {
-		return fmt.Errorf("failed to get path for %v", ip)
-	}
+	objects[object] = true
 
-	_, err = b.s.AddPath(context.Background(), &api.AddPathRequest{
-		Path: p,
-	})
-
-	if err != nil {
-		return err
-	}
-
-	return
+	return nil
 }
 
 // DelHost will inform peers to remove a host
-func (b *Server) DelHost(addr string) (err error) {
+func (b *Server) DelHost(ctx context.Context, addr string, object string) error {
+	b.mtx.Lock()
+	defer b.mtx.Unlock()
+
+	objects, exists := b.tracker[addr]
+	if !exists {
+		log.Debug("[BGP] deleting host - nothing to delete", "addr", addr, "object", object)
+		return nil
+	}
+
 	ip, _, err := net.ParseCIDR(addr)
 	if err != nil {
 		return err
 	}
-	p := b.getPath(ip)
-	if p == nil {
-		return
+
+	delete(objects, object)
+
+	if len(objects) == 0 {
+		p := b.getPath(ip)
+		if p == nil {
+			return nil
+		}
+
+		if err := b.s.DeletePath(apiutil.DeletePathRequest{
+			Paths: []*apiutil.Path{p},
+		}); err != nil {
+			return err
+		}
+		delete(b.tracker, addr)
+		log.Debug("[BGP] deleted host", "addr", addr, "cnt", len(objects), "object", object)
+	} else {
+		log.Debug("[BGP] deleting from tracker only", "addr", addr, "object", object)
 	}
 
-	return b.s.DeletePath(context.Background(), &api.DeletePathRequest{
-		Path: p,
-	})
+	return nil
 }

pkg/bgp/peers.go

@@ -3,24 +3,34 @@ package bgp
 import (
 	"context"
 	"fmt"
+	log "log/slog"
 	"net"
+	"net/netip"
 	"strconv"
-	"strings"
+	"time"
 
-	//nolint
+	"github.com/jpillora/backoff"
+	"github.com/kube-vip/kube-vip/pkg/kubevip"
 	"github.com/kube-vip/kube-vip/pkg/vip"
-	api "github.com/osrg/gobgp/v3/api"
-	"github.com/osrg/gobgp/v3/pkg/server"
-	"google.golang.org/protobuf/types/known/anypb"
+	api "github.com/osrg/gobgp/v4/api"
+
+	"github.com/kube-vip/kube-vip/pkg/utils"
+	"github.com/osrg/gobgp/v4/pkg/apiutil"
+	"github.com/osrg/gobgp/v4/pkg/config/oc"
+	bgp "github.com/osrg/gobgp/v4/pkg/packet/bgp"
+	"github.com/osrg/gobgp/v4/pkg/server"
 )
 
+const defaultBGPPort uint32 = 179
+
 // AddPeer will add peers to the BGP configuration
-func (b *Server) AddPeer(peer Peer) (err error) {
+func (b *Server) AddPeer(ctx context.Context, peer kubevip.BGPPeer) (err error) {
 	p := &api.Peer{
 		Conf: &api.PeerConf{
-			NeighborAddress: peer.Address,
-			PeerAsn:         peer.AS,
-			AuthPassword:    peer.Password,
+			NeighborAddress:   peer.Address,
+			PeerAsn:           peer.AS,
+			NeighborInterface: peer.Interface,
+			AuthPassword:      peer.Password,
 		},
 
 		Timers: &api.Timers{
@@ -40,62 +50,89 @@ func (b *Server) AddPeer(peer Peer) (err error) {
 		Transport: &api.Transport{
 			MtuDiscovery:  true,
 			RemoteAddress: peer.Address,
-			RemotePort:    uint32(179),
+			RemotePort:    defaultBGPPort,
 		},
 	}
 
-	if b.c.MpbgpNexthop != "" {
-		p.AfiSafis = []*api.AfiSafi{
-			{
-				Config: &api.AfiSafiConfig{
-					Family: &api.Family{
-						Afi:  api.Family_AFI_IP,
-						Safi: api.Family_SAFI_UNICAST,
-					},
-					Enabled: true,
-				},
-			},
-			{
-				Config: &api.AfiSafiConfig{
-					Family: &api.Family{
-						Afi:  api.Family_AFI_IP6,
-						Safi: api.Family_SAFI_UNICAST,
-					},
-					Enabled: true,
-				},
-			},
+	if peer.BFDEnabled {
+		p.Bfd = &api.BfdPeerConfig{
+			Enabled:                  true,
+			DesiredMinimumTxInterval: peer.BFDTransmitInterval,
+			RequiredMinimumReceive:   peer.BFDReceiveInterval,
+			DetectionMultiplier:      peer.BFDDetectMultiplier,
+			Port:                     3784, // TODO: Should this be configurable??
 		}
+	}
 
-		peer.setMpbgpOptions(b.c)
-
-		ipv4Address, ipv6Address, err := peer.findMpbgpAddresses(p, b.c)
+	if peer.Interface != "" {
+		neighborAddress, err := getIPv6LinkLocalNeighborAddress(ctx, peer.Interface)
 		if err != nil {
-			return fmt.Errorf("failed to get MP-BGP addresses: %w", err)
+			return fmt.Errorf("failed to get link-local address of interface %s: %w", peer.Interface, err)
 		}
 
-		mask := "128"
-		address := ipv4Address
-		family := api.Family_AFI_IP
-		if vip.IsIPv4(p.Conf.NeighborAddress) {
-			mask = "32"
-			address = ipv6Address
-			family = api.Family_AFI_IP6
+		p.State = &api.PeerState{
+			NeighborAddress: neighborAddress,
 		}
+	}
 
-		err = b.s.AddDefinedSet(context.Background(), &api.AddDefinedSetRequest{
-			DefinedSet: &api.DefinedSet{
-				DefinedType: api.DefinedType_NEIGHBOR,
-				Name:        fmt.Sprintf("peer-%s", p.Conf.NeighborAddress),
-				List:        []string{fmt.Sprintf("%s/%s", p.Conf.NeighborAddress, mask)},
-			},
-		})
+	mpBGP := b.c.MpbgpNexthop
+
+	if peer.MpbgpNexthop != "" {
+		mpBGP = peer.MpbgpNexthop
+	}
+
+	if mpBGP != "" {
+		ipv4Address, ipv6Address, err := peer.FindMpbgpAddresses(p, b.c)
 		if err != nil {
-			return fmt.Errorf("failed to add defined set: %v", err)
-		}
+			log.Error("failed to get MP-BGP addresses, will not us MP-BGP for this host", "error", err)
+		} else {
+			p.AfiSafis = []*api.AfiSafi{
+				{
+					Config: &api.AfiSafiConfig{
+						Family: &api.Family{
+							Afi:  api.Family_AFI_IP,
+							Safi: api.Family_SAFI_UNICAST,
+						},
+						Enabled: true,
+					},
+				},
+				{
+					Config: &api.AfiSafiConfig{
+						Family: &api.Family{
+							Afi:  api.Family_AFI_IP6,
+							Safi: api.Family_SAFI_UNICAST,
+						},
+						Enabled: true,
+					},
+				},
+			}
 
-		if address != "" {
-			if err := insertPolicy(b.s, address, p, family); err != nil {
-				return fmt.Errorf("failed to add policy: %w", err)
+			peer.SetMpbgpOptions(b.c)
+
+			mask := strconv.Itoa(vip.DefaultMaskIPv6)
+			address := ipv4Address
+			family := api.Family_AFI_IP
+			if utils.IsIPv4(p.Conf.NeighborAddress) {
+				mask = strconv.Itoa(vip.DefaultMaskIPv4)
+				address = ipv6Address
+				family = api.Family_AFI_IP6
+			}
+
+			err = b.s.AddDefinedSet(ctx, &api.AddDefinedSetRequest{
+				DefinedSet: &api.DefinedSet{
+					DefinedType: api.DefinedType_DEFINED_TYPE_NEIGHBOR,
+					Name:        fmt.Sprintf("peer-%s", p.Conf.NeighborAddress),
+					List:        []string{fmt.Sprintf("%s/%s", p.Conf.NeighborAddress, mask)},
+				},
+			})
+			if err != nil {
+				return fmt.Errorf("failed to add defined set: %v", err)
+			}
+
+			if address != "" {
+				if err := insertPolicy(ctx, b.s, address, p, family); err != nil {
+					return fmt.Errorf("failed to add policy: %w", err)
+				}
 			}
 		}
 	} else {
@@ -108,168 +145,65 @@ func (b *Server) AddPeer(peer Peer) (err error) {
 		}
 	}
 
-	if err := b.s.AddPeer(context.Background(), &api.AddPeerRequest{Peer: p}); err != nil {
+	if err := b.s.AddPeer(ctx, &api.AddPeerRequest{Peer: p}); err != nil {
 		return fmt.Errorf("failed to add peer: %v", err)
 	}
-
+	log.Info("[BGP]", "peer", p.Conf.NeighborAddress, "AS", p.Conf.PeerAsn, "BFD", p.Bfd)
 	return nil
 }
 
-func (b *Server) getPath(ip net.IP) (path *api.Path) {
+func (b *Server) getPath(ip net.IP) *apiutil.Path {
 	isV6 := ip.To4() == nil
 
-	//nolint
-	originAttr, _ := anypb.New(&api.OriginAttribute{
-		Origin: 0,
-	})
-
 	if !isV6 {
-		//nolint
-		nlri, _ := anypb.New(&api.IPAddressPrefix{
-			Prefix:    ip.String(),
-			PrefixLen: 32,
-		})
-
-		//nolint
-		nhAttr, _ := anypb.New(&api.NextHopAttribute{
-			NextHop: "0.0.0.0", // gobgp will fill this
-		})
-
-		path = &api.Path{
-			Family: &api.Family{
-				Afi:  api.Family_AFI_IP,
-				Safi: api.Family_SAFI_UNICAST,
-			},
-			Nlri:   nlri,
-			Pattrs: []*anypb.Any{originAttr, nhAttr},
-		}
-	} else {
-		//nolint
-		nlri, _ := anypb.New(&api.IPAddressPrefix{
-			Prefix:    ip.String(),
-			PrefixLen: 128,
-		})
-
-		v6Family := &api.Family{
-			Afi:  api.Family_AFI_IP6,
-			Safi: api.Family_SAFI_UNICAST,
-		}
-
-		//nolint
-		mpAttr, _ := anypb.New(&api.MpReachNLRIAttribute{
-			Family:   v6Family,
-			NextHops: []string{"::"}, // gobgp will fill this
-			Nlris:    []*anypb.Any{nlri},
-		})
-
-		path = &api.Path{
-			Family: v6Family,
-			Nlri:   nlri,
-			Pattrs: []*anypb.Any{originAttr, mpAttr},
-		}
-	}
-	return
-}
-
-// ParseBGPPeerConfig - take a string and parses it into an array of peers
-func ParseBGPPeerConfig(config string) (bgpPeers []Peer, err error) {
-	peers := strings.Split(config, ",")
-	if len(peers) == 0 {
-		return nil, fmt.Errorf("no BGP Peer configurations found")
-	}
-
-	for x := range peers {
-		peerStr := peers[x]
-		config := strings.Split(peerStr, "/")
-		peerStr = config[0]
-		if peerStr == "" {
-			continue
-		}
-		isV6Peer := peerStr[0] == '['
-
-		address := ""
-		if isV6Peer {
-			addressEndPos := strings.IndexByte(peerStr, ']')
-			if addressEndPos == -1 {
-				return nil, fmt.Errorf("no matching ] found for IPv6 BGP Peer")
-			}
-			address = peerStr[1:addressEndPos]
-			peerStr = peerStr[addressEndPos+1:]
-		}
-
-		peer := strings.Split(peerStr, ":")
-		if len(peer) < 2 {
-			return nil, fmt.Errorf("mandatory peering params <host>:<AS> incomplete")
-		}
-
-		if !isV6Peer {
-			address = peer[0]
-		}
-
-		ASNumber, err := strconv.ParseUint(peer[1], 10, 32)
+		prefix, err := bgp.NewIPAddrPrefix(netip.MustParsePrefix(
+			fmt.Sprintf("%s/%d", ip.String(), vip.DefaultMaskIPv4),
+		))
 		if err != nil {
-			return nil, fmt.Errorf("BGP Peer AS format error [%s]", peer[1])
+			return nil
 		}
 
-		password := ""
-		if len(peer) >= 3 {
-			password = peer[2]
+		nh, err := bgp.NewPathAttributeNextHop(netip.MustParseAddr("0.0.0.0"))
+		if err != nil {
+			return nil
 		}
 
-		multiHop := false
-		if len(peer) >= 4 {
-			multiHop, err = strconv.ParseBool(peer[3])
-			if err != nil {
-				return nil, fmt.Errorf("BGP MultiHop format error (true/false) [%s]", peer[1])
-			}
+		return &apiutil.Path{
+			Family: bgp.RF_IPv4_UC,
+			Nlri:   prefix,
+			Attrs: []bgp.PathAttributeInterface{
+				bgp.NewPathAttributeOrigin(0),
+				nh,
+			},
 		}
-
-		var port uint64
-		if len(peer) >= 5 {
-			port, err = strconv.ParseUint(peer[4], 10, 16)
-			if err != nil {
-				return nil, fmt.Errorf("BGP Peer AS format error [%s]", peer[1])
-			}
-		} else {
-			port = 179
-		}
-
-		var mpbgpNexthop, mpbgpIPv4, mpbgpIPv6 string
-
-		if len(config) > 1 {
-			configData := strings.Split(config[1], ";")
-			for _, cfg := range configData {
-				c := strings.Split(cfg, "=")
-				switch c[0] {
-				case "mpbgp_nexthop":
-					mpbgpNexthop = c[1]
-				case "mpbgp_ipv4":
-					mpbgpIPv4 = c[1]
-				case "mpbgp_ipv6":
-					mpbgpIPv6 = c[1]
-				default:
-					return nil, fmt.Errorf("peer configuration parameter '%s' is not supported", c[0])
-				}
-			}
-		}
-
-		peerConfig := Peer{
-			Address:      address,
-			AS:           uint32(ASNumber),
-			Port:         uint16(port),
-			Password:     password,
-			MultiHop:     multiHop,
-			MpbgpNexthop: mpbgpNexthop,
-			MpbgpIPv4:    mpbgpIPv4,
-			MpbgpIPv6:    mpbgpIPv6,
-		}
-
-		bgpPeers = append(bgpPeers, peerConfig)
 	}
-	return
-}
 
-func insertPolicy(s *server.BgpServer, address string, p *api.Peer, family api.Family_Afi) error {
+	prefix, err := bgp.NewIPAddrPrefix(netip.MustParsePrefix(
+		fmt.Sprintf("%s/%d", ip.String(), vip.DefaultMaskIPv6),
+	))
+	if err != nil {
+		return nil
+	}
+
+	mpReach, err := bgp.NewPathAttributeMpReachNLRI(
+		bgp.RF_IPv6_UC,
+		[]bgp.PathNLRI{{NLRI: prefix}},
+		netip.MustParseAddr("::"),
+	)
+	if err != nil {
+		return nil
+	}
+
+	return &apiutil.Path{
+		Family: bgp.RF_IPv6_UC,
+		Nlri:   prefix,
+		Attrs: []bgp.PathAttributeInterface{
+			bgp.NewPathAttributeOrigin(0),
+			mpReach,
+		},
+	}
+}
+func insertPolicy(ctx context.Context, s *server.BgpServer, address string, p *api.Peer, family api.Family_Afi) error {
 	familyType := "v4"
 	if family == api.Family_AFI_IP6 {
 		familyType = "v6"
@@ -290,12 +224,12 @@ func insertPolicy(s *server.BgpServer, address string, p *api.Peer, family api.F
 						},
 					},
 					NeighborSet: &api.MatchSet{
-						Type: api.MatchSet_ANY,
+						Type: api.MatchSet_TYPE_ANY,
 						Name: setName,
 					},
 				},
 				Actions: &api.Actions{
-					RouteAction: api.RouteAction_ACCEPT,
+					RouteAction: api.RouteAction_ROUTE_ACTION_ACCEPT,
 					Nexthop: &api.NexthopAction{
 						Address: address,
 					},
@@ -304,28 +238,28 @@ func insertPolicy(s *server.BgpServer, address string, p *api.Peer, family api.F
 			{
 				Conditions: &api.Conditions{
 					NeighborSet: &api.MatchSet{
-						Type: api.MatchSet_ANY,
+						Type: api.MatchSet_TYPE_ANY,
 						Name: setName,
 					},
 				},
 				Actions: &api.Actions{
-					RouteAction: api.RouteAction_ACCEPT,
+					RouteAction: api.RouteAction_ROUTE_ACTION_ACCEPT,
 				},
 			},
 		},
 	}
 
-	err := s.AddPolicy(context.Background(), &api.AddPolicyRequest{
+	err := s.AddPolicy(ctx, &api.AddPolicyRequest{
 		Policy: policy,
 	})
 	if err != nil {
 		return fmt.Errorf("failed to add policy: %w", err)
 	}
 
-	err = s.AddPolicyAssignment(context.Background(), &api.AddPolicyAssignmentRequest{
+	err = s.AddPolicyAssignment(ctx, &api.AddPolicyAssignmentRequest{
 		Assignment: &api.PolicyAssignment{
 			Name:      "global",
-			Direction: api.PolicyDirection_EXPORT,
+			Direction: api.PolicyDirection_POLICY_DIRECTION_EXPORT,
 			Policies: []*api.Policy{
 				{
 					Name: policy.Name,
@@ -339,3 +273,45 @@ func insertPolicy(s *server.BgpServer, address string, p *api.Peer, family api.F
 
 	return nil
 }
+
+func getIPv6LinkLocalNeighborAddress(ctx context.Context, peerInterface string) (string, error) {
+	neighCtx, neighCancel := context.WithTimeout(ctx, time.Minute)
+	defer neighCancel()
+
+	bo := backoff.Backoff{
+		Factor: 2,
+		Jitter: true,
+		Min:    1 * time.Second,
+		Max:    5 * time.Second,
+	}
+
+	maxAttempts := 20.0
+
+	var err error
+	for {
+		select {
+		case <-neighCtx.Done():
+			if err != nil {
+				return "", fmt.Errorf("failed to get link-local address of interface %s: %w", peerInterface, err)
+			}
+			return "", fmt.Errorf("failed to get link-local address of interface %s: %w", peerInterface, neighCtx.Err())
+		default:
+			dur := bo.Duration()
+			var neighborAddress string
+			neighborAddress, err = oc.GetIPv6LinkLocalNeighborAddress(peerInterface)
+			if err != nil && bo.Attempt() >= maxAttempts {
+				return "", fmt.Errorf("failed to get link-local address of interface %s: %w", peerInterface, err)
+			}
+			if neighborAddress != "" {
+				return neighborAddress, nil
+			}
+			t := time.NewTimer(dur)
+			select {
+			case <-neighCtx.Done():
+				t.Stop()
+			case <-t.C:
+			}
+		}
+
+	}
+}

pkg/bgp/peers_test.go

@@ -1,54 +0,0 @@
-package bgp
-
-import (
-	"reflect"
-	"testing"
-)
-
-func TestParseBGPPeerConfig(t *testing.T) {
-	type args struct {
-		config string
-	}
-	tests := []struct {
-		name         string
-		args         args
-		wantBgpPeers []Peer
-		wantErr      bool
-	}{
-		{
-			name: "IPv4, default port",
-			args: args{config: "192.168.0.10:65000::false,192.168.0.11:65000::false"},
-			wantBgpPeers: []Peer{
-				{Address: "192.168.0.10", Port: 179, AS: 65000, MultiHop: false},
-				{Address: "192.168.0.11", Port: 179, AS: 65000, MultiHop: false},
-			},
-		},
-		{
-			name: "IPv4, different port",
-			args: args{config: "192.168.0.10:65000::false:180,192.168.0.11:65000::false:190"},
-			wantBgpPeers: []Peer{
-				{Address: "192.168.0.10", Port: 180, AS: 65000, MultiHop: false},
-				{Address: "192.168.0.11", Port: 190, AS: 65000, MultiHop: false},
-			},
-		},
-		{
-			name: "IPv6, multi-protocol",
-			args: args{config: "[fd00:1111:2222:3333:c7d9:7235:6bf7:5d52]:65501::false/mpbgp_nexthop=auto_sourceif"},
-			wantBgpPeers: []Peer{
-				{Address: "fd00:1111:2222:3333:c7d9:7235:6bf7:5d52", Port: 179, AS: 65501, MultiHop: false, MpbgpNexthop: "auto_sourceif"},
-			},
-		},
-	}
-	for _, tt := range tests {
-		t.Run(tt.name, func(t *testing.T) {
-			gotBgpPeers, err := ParseBGPPeerConfig(tt.args.config)
-			if (err != nil) != tt.wantErr {
-				t.Errorf("ParseBGPPeerConfig() error = %v, wantErr %v", err, tt.wantErr)
-				return
-			}
-			if !reflect.DeepEqual(gotBgpPeers, tt.wantBgpPeers) {
-				t.Errorf("ParseBGPPeerConfig() = %v, want %v", gotBgpPeers, tt.wantBgpPeers)
-			}
-		})
-	}
-}

pkg/bgp/server.go

@@ -3,17 +3,28 @@ package bgp
 import (
 	"context"
 	"fmt"
+	"sync"
 	"time"
 
 	log "log/slog"
 
-	api "github.com/osrg/gobgp/v3/api"
-	gobgp "github.com/osrg/gobgp/v3/pkg/server"
-	"github.com/prometheus/client_golang/prometheus"
+	"github.com/kube-vip/kube-vip/pkg/kubevip"
+	api "github.com/osrg/gobgp/v4/api"
+	"github.com/osrg/gobgp/v4/pkg/apiutil"
+	bgp "github.com/osrg/gobgp/v4/pkg/packet/bgp"
+	gobgp "github.com/osrg/gobgp/v4/pkg/server"
 )
 
+// Server manages a server object
+type Server struct {
+	s       *gobgp.BgpServer
+	c       *kubevip.BGPConfig
+	mtx     sync.Mutex
+	tracker map[string]map[string]bool
+}
+
 // NewBGPServer takes a configuration and returns a running BGP server instance
-func NewBGPServer(c *Config) (b *Server, err error) {
+func NewBGPServer(c kubevip.BGPConfig, logLevel log.Level) (b *Server, err error) {
 	if c.AS == 0 {
 		return nil, fmt.Errorf("you need to provide AS")
 	}
@@ -25,26 +36,23 @@ func NewBGPServer(c *Config) (b *Server, err error) {
 	if len(c.Peers) == 0 {
 		return nil, fmt.Errorf("you need to provide at least one peer")
 	}
+	bgpLogger := log.Default()
+	lvl := &log.LevelVar{}
+	lvl.Set(logLevel)
 
 	b = &Server{
-		s: gobgp.NewBgpServer(),
-		c: c,
-
-		BGPSessionInfoGauge: prometheus.NewGaugeVec(prometheus.GaugeOpts{
-			Namespace: "kube_vip",
-			Subsystem: "manager",
-			Name:      "bgp_session_info",
-			Help:      "Display state of session by setting metric for label value with current state to 1",
-		}, []string{"state", "peer"}),
+		s:       gobgp.NewBgpServer(gobgp.LoggerOption(bgpLogger, lvl)),
+		c:       &c,
+		tracker: make(map[string]map[string]bool),
 	}
 	return
 }
 
 // Start starts the BGP server
-func (b *Server) Start(peerStateChangeCallback func(*api.WatchEventResponse_PeerEvent)) (err error) {
+func (b *Server) Start(ctx context.Context, peerStateChangeCallback func(*apiutil.WatchEventMessage_PeerEvent)) (err error) {
 	go b.s.Serve()
 
-	if err = b.s.StartBgp(context.Background(), &api.StartBgpRequest{
+	if err = b.s.StartBgp(ctx, &api.StartBgpRequest{
 		Global: &api.Global{
 			Asn:        b.c.AS,
 			RouterId:   b.c.RouterID,
@@ -54,25 +62,25 @@ func (b *Server) Start(peerStateChangeCallback func(*api.WatchEventResponse_Peer
 		return
 	}
 
-	if err = b.s.WatchEvent(context.Background(), &api.WatchEventRequest{Peer: &api.WatchEventRequest_Peer{}}, func(r *api.WatchEventResponse) {
-		if p := r.GetPeer(); p != nil && p.Type == api.WatchEventResponse_PeerEvent_STATE {
-			log.Info("[BGP]", "peer", p.String())
+	if err = b.s.WatchEvent(ctx, gobgp.WatchEventMessageCallbacks{
+		OnPeerUpdate: func(p *apiutil.WatchEventMessage_PeerEvent, _ time.Time) {
+			log.Info("[BGP]", "peer", fmt.Sprintf("%+v", p))
 			if peerStateChangeCallback != nil {
 				peerStateChangeCallback(p)
 			}
-		}
-	}); err != nil {
+		},
+	}, gobgp.WatchPeer()); err != nil {
 		return
 	}
 
 	for _, p := range b.c.Peers {
-		if err = b.AddPeer(p); err != nil {
+		if err = b.AddPeer(ctx, p); err != nil {
 			return
 		}
 	}
 
 	if b.c.Zebra.Enabled {
-		if err = b.s.EnableZebra(context.Background(), &api.EnableZebraRequest{
+		if err = b.s.EnableZebra(ctx, &api.EnableZebraRequest{
 			Url:          b.c.Zebra.URL,
 			Version:      b.c.Zebra.Version,
 			SoftwareName: b.c.Zebra.SoftwareName,
@@ -87,7 +95,51 @@ func (b *Server) Start(peerStateChangeCallback func(*api.WatchEventResponse_Peer
 
 // Close will stop a running BGP Server
 func (b *Server) Close() error {
-	ctx, cf := context.WithTimeout(context.Background(), 5*time.Second)
-	defer cf()
+	ctx, cancel := context.WithTimeout(context.Background(), 5*time.Second)
+	defer cancel()
 	return b.s.StopBgp(ctx, &api.StopBgpRequest{})
 }
+
+// ListAdvertisedRoutes retrieves all active routes inside GoBGP's local RIB.
+// It queries the GLOBAL table type to find routes that kube-vip has requested GoBGP to advertise.
+func (b *Server) ListAdvertisedRoutes(ctx context.Context, isIPv6 bool) ([]*api.Destination, error) {
+	afi := bgp.AFI_IP
+
+	if isIPv6 {
+		afi = bgp.AFI_IP6
+	}
+
+	family := bgp.NewFamily(uint16(afi), bgp.SAFI_UNICAST)
+
+	var destinations []*api.Destination
+
+	req := apiutil.ListPathRequest{
+		TableType: api.TableType_TABLE_TYPE_GLOBAL,
+		Family:    family,
+	}
+
+	// GoBGP's embedded server API uses a callback function to stream results
+	// locally without requiring a gRPC client stream setup.
+	err := b.s.ListPath(req, func(prefix bgp.NLRI, paths []*apiutil.Path) {
+		var newPaths []*api.Path
+		for _, p := range paths {
+			np, err := apiutil.NewPath(p.Family, p.Nlri, p.Withdrawal, p.Attrs, time.Unix(p.Age, 0))
+			if err != nil {
+				log.Error("failed to create BGP path details", "err", err)
+				continue
+			}
+			newPaths = append(newPaths, np)
+		}
+		d := &api.Destination{
+			Prefix: prefix.String(),
+			Paths:  newPaths,
+		}
+		destinations = append(destinations, d)
+	})
+
+	if err != nil {
+		return nil, fmt.Errorf("failed to extract local RIB: %w", err)
+	}
+
+	return destinations, nil
+}

pkg/bgp/types.go

@@ -1,145 +0,0 @@
-package bgp
-
-import (
-	"fmt"
-	"net"
-
-	"github.com/kube-vip/kube-vip/pkg/vip"
-	api "github.com/osrg/gobgp/v3/api"
-	gobgp "github.com/osrg/gobgp/v3/pkg/server"
-	"github.com/prometheus/client_golang/prometheus"
-	"github.com/vishvananda/netlink"
-)
-
-// Peer defines a BGP Peer
-type Peer struct {
-	Address      string
-	Port         uint16
-	AS           uint32
-	Password     string
-	MultiHop     bool
-	MpbgpNexthop string
-	MpbgpIPv4    string
-	MpbgpIPv6    string
-}
-
-func (p *Peer) setMpbgpOptions(server *Config) {
-	if p.MpbgpNexthop == "" {
-		p.MpbgpNexthop = server.MpbgpNexthop
-	}
-
-	if p.MpbgpIPv4 == "" {
-		p.MpbgpIPv4 = server.MpbgpIPv4
-	}
-
-	if p.MpbgpIPv6 == "" {
-		p.MpbgpIPv6 = server.MpbgpIPv6
-	}
-}
-
-func (p *Peer) findMpbgpAddresses(ap *api.Peer, server *Config) (string, string, error) {
-	var ipv4Address, ipv6Address string
-	switch p.MpbgpNexthop {
-	case "fixed":
-		ap.Transport.LocalAddress = server.SourceIP
-		if p.MpbgpIPv4 == "" && p.MpbgpIPv6 == "" {
-			return "", "", fmt.Errorf("to use MP-BGP with fixed address at least one IPv4 or IPv6 address has to be provided [current - IPv4: %s, IPv6: %s]",
-				p.MpbgpIPv4, p.MpbgpIPv6)
-		}
-
-		if p.MpbgpIPv4 != "" {
-			if net.ParseIP(p.MpbgpIPv4) == nil {
-				return "", "", fmt.Errorf("provided address '%s' is not a valid IPv4 address", p.MpbgpIPv4)
-			}
-		}
-		if p.MpbgpIPv6 != "" {
-			if net.ParseIP(p.MpbgpIPv6) == nil {
-				return "", "", fmt.Errorf("provided address '%s' is not a valid IPv6 address", p.MpbgpIPv6)
-			}
-		}
-
-		ipv4Address = p.MpbgpIPv4
-		ipv6Address = p.MpbgpIPv6
-	case "auto_sourceip":
-		ap.Transport.LocalAddress = server.SourceIP
-
-		// Resolve the local interface by SourceIP
-		iface, err := vip.GetInterfaceByIP(server.SourceIP)
-		if err != nil {
-			return "", "", fmt.Errorf("failed to get interface by IP: %v", err)
-		}
-
-		if vip.IsIPv4(server.SourceIP) {
-			// Get the non link-local IPv6 address on that interface
-			ipv6Address, err = vip.GetNonLinkLocalIP(iface, netlink.FAMILY_V6)
-			if err != nil {
-				return "", "", fmt.Errorf("failed to get non link-local IPv6 address: %v", err)
-			}
-		} else {
-			// Get the non link-local IPv4 address on that interface
-			ipv4Address, err = vip.GetNonLinkLocalIP(iface, netlink.FAMILY_V4)
-			if err != nil {
-				return "", "", fmt.Errorf("failed to get non link-local IPv4 address: %v", err)
-			}
-		}
-	case "auto_sourceif":
-		ap.Transport.BindInterface = server.SourceIF
-
-		iface, err := netlink.LinkByName(server.SourceIF)
-		if err != nil {
-			return "", "", fmt.Errorf("failed to get interface by name: %v", err)
-		}
-
-		// Get the non link-local IPv4 address on that interface
-		ipv4Address, err = vip.GetNonLinkLocalIP(&iface, netlink.FAMILY_V4)
-		if err != nil {
-			return "", "", fmt.Errorf("failed to get non link-local IPv4 address: %v", err)
-		}
-
-		// Get the non link-local IPv6 address on that interface
-		ipv6Address, err = vip.GetNonLinkLocalIP(&iface, netlink.FAMILY_V6)
-		if err != nil {
-			return "", "", fmt.Errorf("failed to get non link-local IPv6 address: %v", err)
-		}
-	default:
-		return "", "", fmt.Errorf("option %s for MP-BPG nexthop is not supported", server.MpbgpNexthop)
-	}
-
-	return ipv4Address, ipv6Address, nil
-}
-
-// Config defines the BGP server configuration
-type Config struct {
-	AS           uint32
-	RouterID     string
-	SourceIP     string
-	SourceIF     string
-	MpbgpNexthop string
-	MpbgpIPv4    string
-	MpbgpIPv6    string
-
-	HoldTime          uint64
-	KeepaliveInterval uint64
-
-	Peers []Peer
-
-	Zebra ZebraConfig
-}
-
-// Defines Zebra connection configuration. More on the topic - https://github.com/osrg/gobgp/blob/master/docs/sources/zebra.md#configuration
-type ZebraConfig struct {
-	Enabled      bool
-	URL          string
-	Version      uint32
-	SoftwareName string
-}
-
-// Server manages a server object
-type Server struct {
-	s *gobgp.BgpServer
-	c *Config
-
-	// This is a prometheus gauge indicating the state of the sessions.
-	// 1 means "ESTABLISHED", 0 means "NOT ESTABLISHED"
-	BGPSessionInfoGauge *prometheus.GaugeVec
-}

pkg/cluster/cluster.go

@@ -1,28 +1,40 @@
 package cluster
 
 import (
-	"sync"
+	"crypto/tls"
+	"crypto/x509"
+	"fmt"
+	"net/http"
+	"os"
+	"time"
 
 	log "log/slog"
 
 	"github.com/kube-vip/kube-vip/pkg/arp"
 	"github.com/kube-vip/kube-vip/pkg/kubevip"
 	"github.com/kube-vip/kube-vip/pkg/networkinterface"
+	"github.com/kube-vip/kube-vip/pkg/node"
+	"github.com/kube-vip/kube-vip/pkg/route"
+	"github.com/kube-vip/kube-vip/pkg/utils"
 	"github.com/kube-vip/kube-vip/pkg/vip"
 )
 
 // Cluster - The Cluster object manages the state of the cluster for a particular node
 type Cluster struct {
-	stop      chan bool
-	completed chan bool
-	once      sync.Once
-	Network   []vip.Network
-	arpMgr    *arp.Manager
+	stop                  chan bool
+	Network               []vip.Network
+	arpMgr                *arp.Manager
+	routeMgr              *route.Manager
+	nodeLabelMgr          node.Labeler
+	labelAdded            bool
+	healthCheckHTTPClient *http.Client
 }
 
 // InitCluster - Will attempt to initialise all of the required settings for the cluster
-func InitCluster(c *kubevip.Config, disableVIP bool, intfMgr *networkinterface.Manager, arpMgr *arp.Manager) (*Cluster, error) {
+func InitCluster(c *kubevip.Config, disableVIP bool, intfMgr *networkinterface.Manager, arpMgr *arp.Manager,
+	routeMgr *route.Manager, nodeLabelMgr node.Labeler) (*Cluster, error) {
 	var networks []vip.Network
+	var healthCheckHTTPClient *http.Client
 	var err error
 
 	if !disableVIP {
@@ -32,10 +44,22 @@ func InitCluster(c *kubevip.Config, disableVIP bool, intfMgr *networkinterface.M
 			return nil, err
 		}
 	}
+
+	if c.ControlPlaneHealthCheck.Address != "" {
+		healthCheckHTTPClient, err = newHealthCheckHTTPClient(c)
+		if err != nil {
+			return nil, fmt.Errorf("initializing BGP health check client: %w", err)
+		}
+	}
+
 	// Initialise the Cluster structure
 	newCluster := &Cluster{
-		Network: networks,
-		arpMgr:  arpMgr,
+		Network:               networks,
+		arpMgr:                arpMgr,
+		stop:                  make(chan bool),
+		routeMgr:              routeMgr,
+		nodeLabelMgr:          nodeLabelMgr,
+		healthCheckHTTPClient: healthCheckHTTPClient,
 	}
 
 	log.Debug("service security", "enabled", c.EnableServiceSecurity)
@@ -54,9 +78,10 @@ func startNetworking(c *kubevip.Config, intfMgr *networkinterface.Manager) ([]vi
 
 	networks := []vip.Network{}
 	for _, addr := range addresses {
-		network, err := vip.NewConfig(addr, c.Interface, c.LoInterfaceGlobalScope, c.VIPSubnet, c.DDNS, c.RoutingTableID,
-			c.RoutingTableType, c.RoutingProtocol, c.DNSMode, c.LoadBalancerForwardingMethod, c.IptablesBackend,
-			c.EnableLoadBalancer, c.EnableServiceSecurity, intfMgr)
+		network, err := vip.NewConfig(addr, c.Interface, c.LoInterfaceGlobalScope, c.VIPSubnet, c.DDNS, c.DHCPMode,
+			c.RequireDualStack, c.IsDualStack, c.RoutingTableID, c.RoutingTableType, c.RoutingProtocol, c.DNSMode,
+			c.LoadBalancerForwardingMethod, c.IptablesBackend, c.EnableLoadBalancer, c.LoadBalancerPort,
+			c.EnableServiceSecurity, intfMgr, c.EgressWithNftables, c.SkipDAD)
 		if err != nil {
 			return nil, err
 		}
@@ -70,11 +95,78 @@ func startNetworking(c *kubevip.Config, intfMgr *networkinterface.Manager) ([]vi
 func (cluster *Cluster) Stop() {
 	// Close the stop channel, which will shut down the VIP (if needed)
 	if cluster.stop != nil {
-		cluster.once.Do(func() { // Ensure that the close channel can only ever be called once
-			close(cluster.stop)
-		})
+		close(cluster.stop)
+		cluster.stop = make(chan bool) // recreate channel for future use
+	}
+}
+
+func newHealthCheckHTTPClient(c *kubevip.Config) (*http.Client, error) {
+	defaultTransport, ok := http.DefaultTransport.(*http.Transport)
+	if !ok {
+		return nil, fmt.Errorf("unexpected default HTTP transport type %T", http.DefaultTransport)
 	}
 
-	// Wait until the completed channel is closed, signallign all shutdown tasks completed
-	<-cluster.completed
+	transport := defaultTransport.Clone()
+	if c.ControlPlaneHealthCheck.CAPath != "" {
+		caCert, err := os.ReadFile(c.ControlPlaneHealthCheck.CAPath)
+		if err != nil {
+			return nil, fmt.Errorf("reading health check CA cert %q: %w", c.ControlPlaneHealthCheck.CAPath, err)
+		}
+
+		rootCAs, err := x509.SystemCertPool()
+		if err != nil || rootCAs == nil {
+			rootCAs = x509.NewCertPool()
+		}
+		if !rootCAs.AppendCertsFromPEM(caCert) {
+			return nil, fmt.Errorf("health check CA cert %q contains no valid certificates", c.ControlPlaneHealthCheck.CAPath)
+		}
+
+		tlsConfig := &tls.Config{MinVersion: tls.VersionTLS12}
+		if transport.TLSClientConfig != nil {
+			tlsConfig = transport.TLSClientConfig.Clone()
+		}
+		tlsConfig.RootCAs = rootCAs
+		transport.TLSClientConfig = tlsConfig
+	}
+
+	return &http.Client{
+		Timeout:   time.Duration(c.ControlPlaneHealthCheck.TimeoutSeconds) * time.Second,
+		Transport: transport,
+	}, nil
+}
+
+// cleanupVIPs handles VIP removal based on the PreserveVIPOnLeadershipLoss configuration.
+// When preservation is enabled, IPv6 VIPs are always removed immediately to prevent DAD
+// failures on the new leader, while IPv4 VIPs are intentionally left in place.
+// When preservation is disabled (legacy behavior), all VIPs are removed.
+func (cluster *Cluster) cleanupVIPs(c *kubevip.Config) {
+	for i := range cluster.Network {
+		if c.EnableARP && cluster.arpMgr.Count(cluster.Network[i].ARPName()) > 1 {
+			continue
+		}
+
+		if c.PreserveVIPOnLeadershipLoss {
+			if utils.IsIPv6(cluster.Network[i].IP()) {
+				log.Info("[VIP] Removing IPv6 VIP immediately (required to prevent DAD failures on new leader)", "ip", cluster.Network[i].IP())
+				deleted, err := cluster.Network[i].DeleteIP()
+				if err != nil {
+					log.Warn(err.Error())
+				}
+				if deleted {
+					log.Info("deleted address", "IP", cluster.Network[i].IP(), "interface", cluster.Network[i].Interface())
+				}
+			} else {
+				log.Info("[VIP] Preserving IPv4 VIP address on interface, only stopped ARP broadcasting", "ip", cluster.Network[i].IP())
+			}
+		} else {
+			log.Info("[VIP] Deleting VIP", "ip", cluster.Network[i].IP())
+			deleted, err := cluster.Network[i].DeleteIP()
+			if err != nil {
+				log.Warn(err.Error())
+			}
+			if deleted {
+				log.Info("deleted address", "IP", cluster.Network[i].IP(), "interface", cluster.Network[i].Interface())
+			}
+		}
+	}
 }

pkg/cluster/clusterDDNS.go

@@ -2,6 +2,7 @@ package cluster
 
 import (
 	"context"
+	"sync"
 
 	"github.com/kube-vip/kube-vip/pkg/vip"
 )
@@ -12,16 +13,14 @@ import (
 // during runtime if IP changes, startDDNS don't have to do reconfigure because
 // dnsUpdater already have the functionality to keep trying resolve the IP
 // and update the VIP configuration if it changes
-func (cluster *Cluster) StartDDNS(ctx context.Context) error {
-	for i := range cluster.Network {
-		ddnsMgr := vip.NewDDNSManager(ctx, cluster.Network[i])
-		ip, err := ddnsMgr.Start()
-		if err != nil {
-			return err
-		}
-		if err = cluster.Network[i].SetIP(ip); err != nil {
-			return err
-		}
+func (cluster *Cluster) StartDDNS(ctx context.Context, network vip.Network, backoffAttempts uint, wg *sync.WaitGroup) error {
+	ddnsMgr := vip.NewDDNSManager(network, backoffAttempts)
+	ip, err := ddnsMgr.Start(ctx, wg)
+	if err != nil {
+		return err
+	}
+	if err = network.SetIP(ip); err != nil {
+		return err
 	}
 
 	return nil

pkg/cluster/clusterLeaderElection.go

@@ -3,131 +3,59 @@ package cluster
 import (
 	"context"
 	"fmt"
-	"os"
-	"os/signal"
-	"path/filepath"
-	"syscall"
-	"time"
+	"sync"
 
-	"github.com/davecgh/go-spew/spew"
 	"github.com/kube-vip/kube-vip/pkg/bgp"
-	"github.com/kube-vip/kube-vip/pkg/etcd"
-	"github.com/kube-vip/kube-vip/pkg/k8s"
+	"github.com/kube-vip/kube-vip/pkg/election"
 	"github.com/kube-vip/kube-vip/pkg/kubevip"
-	"github.com/kube-vip/kube-vip/pkg/loadbalancer"
+	"github.com/kube-vip/kube-vip/pkg/lease"
+	"github.com/kube-vip/kube-vip/pkg/utils"
 
 	log "log/slog"
-
-	clientv3 "go.etcd.io/etcd/client/v3"
-	v1 "k8s.io/api/core/v1"
-	apierrors "k8s.io/apimachinery/pkg/api/errors"
-	metav1 "k8s.io/apimachinery/pkg/apis/meta/v1"
-	"k8s.io/apimachinery/pkg/watch"
-	"k8s.io/client-go/kubernetes"
-	"k8s.io/client-go/tools/cache"
-	"k8s.io/client-go/tools/leaderelection"
-	"k8s.io/client-go/tools/leaderelection/resourcelock"
-	watchtools "k8s.io/client-go/tools/watch"
 )
 
-// Manager degines the manager of the load-balancing services
-type Manager struct {
-	KubernetesClient   *kubernetes.Clientset
-	RetryWatcherClient *kubernetes.Clientset
-	// This channel is used to signal a shutdown
-	SignalChan chan os.Signal
+// StartCluster - Begins a running instance of the Leader Election cluster
+func (cluster *Cluster) StartCluster(ctx context.Context, c *kubevip.Config,
+	em *election.Manager, bgpServer *bgp.Server, leaseMgr *lease.Manager, killFunc func()) error {
 
-	EtcdClient *clientv3.Client
-}
+	ns, leaseName := lease.NamespaceName(c.LeaseName, c)
 
-// NewManager will create a new managing object
-func NewManager(path string, inCluster bool, port int) (*Manager, error) {
-	var hostname string
+	leaseID := lease.NewID(c.LeaderElectionType, ns, leaseName)
 
-	// If inCluster is set then it will likely have started as a static pod or won't have the
-	// VIP up before trying to connect to the API server, we set the API endpoint to this machine to
-	// ensure connectivity. Else if the path passed is empty and not running in the cluster,
-	// attempt to look for a kubeconfig in the default HOME dir.
+	log.Info("cluster membership", "namespace", leaseID.Namespace(), "lock", leaseID.Name(), "id", c.NodeName)
 
-	hostname = fmt.Sprintf("kubernetes:%v", port)
+	objectName := lease.ObjectName(leaseID, "cp")
+	objLease := leaseMgr.Add(ctx, leaseID)
+	isNew := objLease.Add(objectName)
 
-	if len(path) == 0 && !inCluster {
-		path = filepath.Join(os.Getenv("HOME"), ".kube", "config")
+	wg := sync.WaitGroup{}
+	defer wg.Wait()
 
-		// We modify the config so that we can always speak to the correct host
-		id, err := os.Hostname()
-		if err != nil {
-			return nil, err
+	// Start a goroutine that will delete the lease when the service context is cancelled.
+	// This is important for proper cleanup when a service is deleted - it ensures that
+	// the lease context (svcLease.Ctx) gets cancelled, which causes RunOrDie to return.
+	// Without this, RunOrDie would continue running until leadership is naturally lost.
+	wg.Go(func() {
+		<-objLease.Ctx.Done()
+		leaseMgr.Delete(leaseID, objectName, objLease)
+	})
+
+	if !isNew {
+		log.Debug("this election was already done, waiting for it to finish", "lease", leaseName)
+		<-objLease.Ctx.Done()
+		return nil
+	}
+
+	wg.Go(func() {
+		select {
+		case <-cluster.stop:
+		case <-ctx.Done():
 		}
 
-		hostname = fmt.Sprintf("%s:%v", id, port)
-	}
-
-	config, err := k8s.NewRestConfig(path, inCluster, hostname)
-	if err != nil {
-		return nil, fmt.Errorf("failed to create k8s REST config: %w", err)
-	}
-
-	clientset, err := k8s.NewClientset(config)
-	if err != nil {
-		return nil, fmt.Errorf("error creating a new k8s clientset: %v", err)
-	}
-
-	rwConfig, err := k8s.NewRestConfig(path, inCluster, hostname)
-	if err != nil {
-		return nil, fmt.Errorf("failed to create k8s REST config for retryClientSet: %w", err)
-	}
-
-	rwConfig.Timeout = 0 // empty value to disable the timeout
-	rwClientSet, err := k8s.NewClientset(rwConfig)
-	if err != nil {
-		return nil, fmt.Errorf("failed to create k8s client for retry watcher: %w", err)
-	}
-
-	return &Manager{
-		KubernetesClient:   clientset,
-		RetryWatcherClient: rwClientSet,
-	}, nil
-}
-
-// StartCluster - Begins a running instance of the Leader Election cluster
-func (cluster *Cluster) StartCluster(c *kubevip.Config, sm *Manager, bgpServer *bgp.Server) error {
-	var err error
-
-	log.Info("cluster membership", "namespace", c.Namespace, "lock", c.LeaseName, "id", c.NodeName)
-
-	// use a Go context so we can tell the leaderelection code when we
-	// want to step down
-	ctx, cancel := context.WithCancel(context.Background())
-	defer cancel()
-
-	// use a Go context so we can tell the arp loop code when we
-	// want to step down
-	ctxArp, cancelArp := context.WithCancel(context.Background())
-	defer cancelArp()
-
-	// use a Go context so we can tell the dns loop code when we
-	// want to step down
-	ctxDNS, cancelDNS := context.WithCancel(context.Background())
-	defer cancelDNS()
-
-	// listen for interrupts or the Linux SIGTERM signal and cancel
-	// our context, which the leader election code will observe and
-	// step down
-	signalChan := make(chan os.Signal, 1)
-	// Add Notification for Userland interrupt
-	signal.Notify(signalChan, syscall.SIGINT)
-
-	// Add Notification for SIGTERM (sent from Kubernetes)
-	signal.Notify(signalChan, syscall.SIGTERM)
-
-	go func() {
-		<-signalChan
 		log.Info("Received termination, signaling cluster shutdown")
-		// Cancel the context, which will in turn cancel the leadership
-		cancel()
-		// Cancel the arp context, which will in turn stop any broadcasts
-	}()
+		// Cancel the leader context, which will in turn cancel the leadership
+		objLease.Cancel()
+	})
 
 	// (attempt to) Remove the virtual IP, in case it already exists
 
@@ -141,250 +69,147 @@ func (cluster *Cluster) StartCluster(c *kubevip.Config, sm *Manager, bgpServer *
 		}
 	}
 
-	// Defer a function to check if the bgpServer has been created and if so attempt to close it
+	objLease.Lock()
+
 	defer func() {
-		if bgpServer != nil {
-			bgpServer.Close()
-		}
+		objLease.Unlock()
 	}()
 
-	if c.EnableBGP && bgpServer == nil {
-		// Lets start BGP
-		log.Info("Starting the BGP server to advertise VIP routes to VGP peers")
-		bgpServer, err = bgp.NewBGPServer(&c.BGPConfig)
-		if err != nil {
-			log.Error("new BGP server", "err", err)
-		}
-		if err := bgpServer.Start(nil); err != nil {
-			log.Error("starting BGP server", "err", err)
+	// this object is sharing lease with another object
+	if objLease.Elected.Load() {
+		log.Debug("this election was already done, shared lease", "lease", leaseName)
+		// wait for leader election to start or context to be done
+		select {
+		case <-objLease.Started:
+		case <-objLease.Ctx.Done():
+			// Lease was cancelled (e.g., leader election ended), return immediately
+			// This allows the restart loop to create a fresh lease
+			log.Debug("lease context cancelled before leader election started", "lease", leaseName)
+			return fmt.Errorf("lease %q context cancelled before leader election started", leaseName)
 		}
+
+		cluster.OnStartedLeading(c, objLease, em, bgpServer, killFunc, true)
+
+		log.Debug("cluster waiting for leader context done", "lease", leaseName)
+		// wait for leaderelection to be finished
+		<-objLease.Ctx.Done()
+
+		cluster.OnStoppedLeading(c, objLease, bgpServer)
+
+		return nil
 	}
 
-	run := &runConfig{
-		config:  c,
-		leaseID: c.NodeName,
-		sm:      sm,
-		onStartedLeading: func(ctx context.Context) { //nolint TODO: potential clean code
-			// As we're leading lets start the vip service
-			err := cluster.vipService(ctxArp, ctxDNS, c, sm, bgpServer, cancel)
+	run := &election.RunConfig{
+		Config:           c,
+		LeaseID:          leaseID,
+		LeaseAnnotations: c.LeaseAnnotations,
+		Mgr:              em,
+		OnStartedLeading: func(context.Context) { //nolint TODO: potential clean code
+			cluster.OnStartedLeading(c, objLease, em, bgpServer, killFunc, false)
+		},
+		OnStoppedLeading: func() {
+			objLease.Elected.Store(false)
+			cluster.OnStoppedLeading(c, objLease, bgpServer)
+		},
+		OnNewLeader: func(identity string) {
+			cluster.OnNewLeader(identity, c)
+		},
+	}
+
+	if err := election.RunOrDie(objLease.Ctx, run, c); err != nil {
+		cluster.Stop()
+		return fmt.Errorf("leaderelection failed: %w", err)
+	}
+
+	return nil
+}
+
+func (cluster *Cluster) OnStartedLeading(c *kubevip.Config, objLease *lease.Lease,
+	em *election.Manager, bgpServer *bgp.Server, killFunc func(), isShared bool) {
+	objLease.Elected.Store(true)
+	objLease.Unlock()
+
+	// When we become leader, ensure we can take over VIPs even if they're preserved on other nodes
+	if !isShared {
+		close(objLease.Started)
+	}
+
+	labels := generateLabelsFromConfig(c.Address, kubevip.HasIP)
+	if err := cluster.nodeLabelMgr.AddLabel(labels); err != nil {
+		log.Error("error adding label to node", "err", err)
+	}
+	cluster.labelAdded = true
+
+	if c.PreserveVIPOnLeadershipLoss {
+		log.Info("Becoming leader with VIP preservation enabled - ensuring VIP takeover")
+		// Force add the VIPs (this will work even if they exist due to the precheck logic)
+		for i := range cluster.Network {
+			added, err := cluster.Network[i].AddIP(true, false)
 			if err != nil {
-				log.Error("starting VIP service on leader", "err", err)
+				log.Error("failed to ensure VIP on leader takeover", "vip", cluster.Network[i].IP(), "err", err)
+			} else if added {
+				log.Info("took over VIP as new leader", "IP", cluster.Network[i].IP(), "interface", cluster.Network[i].Interface())
+			} else {
+				log.Info("VIP already configured on interface", "IP", cluster.Network[i].IP(), "interface", cluster.Network[i].Interface())
 			}
-		},
-		onStoppedLeading: func() {
-			// we can do cleanup here
-			log.Info("This node is becoming a follower within the cluster")
-
-			// Stop the dns context
-			cancelDNS()
-			// Stop the Arp context if it is running
-			cancelArp()
-
-			// Stop the BGP server
-			if bgpServer != nil {
-				err := bgpServer.Close()
-				if err != nil {
-					log.Warn("close BGP server", "err", err)
-				}
-			}
-
-			for i := range cluster.Network {
-				deleted, err := cluster.Network[i].DeleteIP()
-				if err != nil {
-					log.Warn("delete VIP", "err", err)
-				}
-				if deleted {
-					log.Info("deleted address", "IP", cluster.Network[i].IP(), "interface", cluster.Network[i].Interface())
-				}
-			}
-
-			log.Error("lost leadership, restarting kube-vip")
-			panic("") // TODO - we could also return here
-		},
-		onNewLeader: func(identity string) {
-			// we're notified when new leader elected
-			log.Info("New leader", "leader", identity)
-		},
+		}
 	}
 
-	switch c.LeaderElectionType {
-	case "kubernetes", "":
-		cluster.runKubernetesLeaderElectionOrDie(ctx, run)
-	case "etcd":
-		cluster.runEtcdLeaderElectionOrDie(ctx, run)
-	default:
-		log.Info(fmt.Sprintf("LeaderElectionMode %s not supported, exiting", c.LeaderElectionType))
-	}
-
-	return nil
-}
-
-type runConfig struct {
-	config  *kubevip.Config
-	leaseID string
-	sm      *Manager
-
-	// onStartedLeading is called when this member starts leading.
-	onStartedLeading func(context.Context)
-	// onStoppedLeading is called when this member stops leading.
-	onStoppedLeading func()
-	// onNewLeader is called when the client observes a leader that is
-	// not the previously observed leader. This includes the first observed
-	// leader when the client starts.
-	onNewLeader func(identity string)
-}
-
-func (cluster *Cluster) runKubernetesLeaderElectionOrDie(ctx context.Context, run *runConfig) {
-	// we use the Lease lock type since edits to Leases are less common
-	// and fewer objects in the cluster watch "all Leases".
-	lock := &resourcelock.LeaseLock{
-		LeaseMeta: metav1.ObjectMeta{
-			Name:        run.config.LeaseName,
-			Namespace:   run.config.Namespace,
-			Annotations: run.config.LeaseAnnotations,
-		},
-		Client: run.sm.KubernetesClient.CoordinationV1(),
-		LockConfig: resourcelock.ResourceLockConfig{
-			Identity: run.leaseID,
-		},
-	}
-
-	// start the leader election code loop
-	leaderelection.RunOrDie(ctx, leaderelection.LeaderElectionConfig{
-		Lock: lock,
-		// IMPORTANT: you MUST ensure that any code you have that
-		// is protected by the lease must terminate **before**
-		// you call cancel. Otherwise, you could have a background
-		// loop still running and another process could
-		// get elected before your background loop finished, violating
-		// the stated goal of the lease.
-		ReleaseOnCancel: true,
-		LeaseDuration:   time.Duration(run.config.LeaseDuration) * time.Second,
-		RenewDeadline:   time.Duration(run.config.RenewDeadline) * time.Second,
-		RetryPeriod:     time.Duration(run.config.RetryPeriod) * time.Second,
-		Callbacks: leaderelection.LeaderCallbacks{
-			OnStartedLeading: run.onStartedLeading,
-			OnStoppedLeading: run.onStoppedLeading,
-			OnNewLeader:      run.onNewLeader,
-		},
-	})
-}
-
-func (cluster *Cluster) runEtcdLeaderElectionOrDie(ctx context.Context, run *runConfig) {
-	etcd.RunElectionOrDie(ctx, &etcd.LeaderElectionConfig{
-		EtcdConfig:           etcd.ClientConfig{Client: run.sm.EtcdClient},
-		Name:                 run.config.LeaseName,
-		MemberID:             run.leaseID,
-		LeaseDurationSeconds: int64(run.config.LeaseDuration),
-		Callbacks: etcd.LeaderCallbacks{
-			OnStartedLeading: run.onStartedLeading,
-			OnStoppedLeading: run.onStoppedLeading,
-			OnNewLeader:      run.onNewLeader,
-		},
-	})
-}
-
-func (sm *Manager) NodeWatcher(ctxArp context.Context, lb *loadbalancer.IPVSLoadBalancer, port uint16) error {
-	// Use a restartable watcher, as this should help in the event of etcd or timeout issues
-	log.Info("Kube-Vip is watching nodes for control-plane labels")
-
-	listOptions := metav1.ListOptions{
-		LabelSelector: "node-role.kubernetes.io/control-plane",
-	}
-
-	rw, err := watchtools.NewRetryWatcherWithContext(ctxArp, "1", &cache.ListWatch{
-		WatchFunc: func(_ metav1.ListOptions) (watch.Interface, error) {
-			return sm.RetryWatcherClient.CoreV1().Nodes().Watch(context.Background(), listOptions)
-		},
-	})
+	// As we're leading lets start the vip service
+	err := cluster.StartVipService(objLease.Ctx, c, em, bgpServer, killFunc)
 	if err != nil {
-		return fmt.Errorf("error creating label watcher: %s", err.Error())
+		log.Error("starting VIP service on leader", "err", err)
+		killFunc()
 	}
-
-	go func() {
-		<-sm.SignalChan
-		log.Info("Received termination, signaling shutdown")
-		// Cancel the context
-		rw.Stop()
-	}()
-
-	ch := rw.ResultChan()
-	// defer rw.Stop()
-
-	for event := range ch {
-		// We need to inspect the event and get ResourceVersion out of it
-		switch event.Type {
-		case watch.Added, watch.Modified:
-			node, ok := event.Object.(*v1.Node)
-			if !ok {
-				return fmt.Errorf("unable to parse Kubernetes Node from Annotation watcher")
-			}
-			// Find the node IP address (this isn't foolproof)
-			for x := range node.Status.Addresses {
-				if node.Status.Addresses[x].Type == v1.NodeInternalIP {
-					if checkIfNodeIsReady(node) {
-						err = lb.AddBackend(node.Status.Addresses[x].Address, port)
-						if err != nil {
-							log.Error("add IPVS backend", "err", err)
-						}
-					} else {
-						err = lb.RemoveBackend(node.Status.Addresses[x].Address, port)
-						if err != nil {
-							log.Error("remove IPVS backend", "err", err)
-						}
-					}
-				}
-			}
-		case watch.Deleted:
-			node, ok := event.Object.(*v1.Node)
-			if !ok {
-				return fmt.Errorf("unable to parse Kubernetes Node from Annotation watcher")
-			}
-
-			// Find the node IP address (this isn't foolproof)
-			for x := range node.Status.Addresses {
-				if node.Status.Addresses[x].Type == v1.NodeInternalIP {
-					err = lb.RemoveBackend(node.Status.Addresses[x].Address, port)
-					if err != nil {
-						log.Error("Del IPVS backend", "err", err)
-					}
-				}
-			}
-
-			log.Info("Node deleted", "name", node.Name)
-
-		case watch.Bookmark:
-			// Un-used
-		case watch.Error:
-			log.Error("Error attempting to watch Kubernetes Nodes")
-
-			// This round trip allows us to handle unstructured status
-			errObject := apierrors.FromObject(event.Object)
-			statusErr, ok := errObject.(*apierrors.StatusError)
-			if !ok {
-				log.Error(spew.Sprintf("Received an error which is not *metav1.Status but %#+v", event.Object))
-			}
-
-			status := statusErr.ErrStatus
-			log.Error("watcher", "status", status)
-		default:
-		}
-	}
-
-	log.Info("Exiting Node watcher")
-	return nil
 }
 
-func checkIfNodeIsReady(node *v1.Node) bool {
-	if node == nil {
-		return false
+func (cluster *Cluster) OnStoppedLeading(c *kubevip.Config, objLease *lease.Lease,
+	bgpServer *bgp.Server) {
+	// we can do cleanup here
+	log.Info("This node is becoming a follower within the cluster")
+
+	if cluster.labelAdded {
+		labels := generateLabelsFromConfig(c.Address, kubevip.HasIP)
+		if err := cluster.nodeLabelMgr.RemoveLabel(labels); err != nil {
+			log.Error("error removing label from node", "err", err)
+		}
+		cluster.labelAdded = false
 	}
-	for _, condition := range node.Status.Conditions {
-		if condition.Type == v1.NodeReady {
-			if condition.Status == v1.ConditionTrue {
-				return true
+
+	// Stop the cluster context if it is running
+	objLease.Cancel()
+
+	cluster.cleanupVIPs(c)
+
+	log.Error("lost leadership, restarting kube-vip")
+}
+
+func (cluster *Cluster) OnNewLeader(identity string, c *kubevip.Config) {
+	// we're notified when new leader elected
+	log.Info("New leader", "leader", identity)
+
+	// If we're not the new leader and we have VIPs preserved from previous leadership,
+	// we need to clean them up to avoid conflicts.
+	if identity != c.NodeName && c.PreserveVIPOnLeadershipLoss {
+		log.Info("Cleaning up preserved VIPs as another node became leader", "new_leader", identity)
+		for i := range cluster.Network {
+			deleted, err := cluster.Network[i].DeleteIP()
+			if err != nil {
+				log.Warn("failed to cleanup preserved VIP", "vip", cluster.Network[i].IP(), "err", err)
+			}
+			if deleted {
+				log.Info("cleaned up preserved VIP to avoid conflict", "IP", cluster.Network[i].IP(),
+					"interface", cluster.Network[i].Interface(), "new_leader", identity)
+			} else {
+				log.Debug("VIP was not present on this node", "IP", cluster.Network[i].IP(),
+					"interface", cluster.Network[i].Interface())
 			}
 		}
 	}
-	return false
+}
+
+func generateLabelsFromConfig(addr, labelKey string) map[string]string {
+	return map[string]string{
+		labelKey: utils.SanitizeIPForLabel(addr),
+	}
 }

pkg/cluster/cluster_test.go

@@ -0,0 +1,125 @@
+package cluster_test
+
+import (
+	"crypto/ecdsa"
+	"crypto/elliptic"
+	"crypto/rand"
+	"crypto/x509"
+	"encoding/pem"
+	"math/big"
+	"os"
+	"path/filepath"
+	"strings"
+	"testing"
+	"time"
+
+	"github.com/kube-vip/kube-vip/pkg/cluster"
+	"github.com/kube-vip/kube-vip/pkg/kubevip"
+)
+
+func TestInitCluster_HealthCheckClientNoCA(t *testing.T) {
+	t.Parallel()
+	cfg := &kubevip.Config{
+		EnableBGP: true,
+		ControlPlaneHealthCheck: kubevip.HealthCheck{
+			Address:        "http://localhost:6443/livez",
+			TimeoutSeconds: 5,
+		},
+	}
+
+	_, err := cluster.InitCluster(cfg, true, nil, nil, nil, nil)
+	if err != nil {
+		t.Fatalf("unexpected error: %v", err)
+	}
+}
+
+func TestInitCluster_HealthCheckClientValidCA(t *testing.T) {
+	t.Parallel()
+	caPEM := generateTestCACert(t)
+	caFile := filepath.Join(t.TempDir(), "ca.crt")
+	if err := os.WriteFile(caFile, caPEM, 0o600); err != nil {
+		t.Fatal(err)
+	}
+
+	cfg := &kubevip.Config{
+		EnableBGP: true,
+		ControlPlaneHealthCheck: kubevip.HealthCheck{
+			Address:        "https://localhost:6443/livez",
+			TimeoutSeconds: 3,
+			CAPath:         caFile,
+		},
+	}
+
+	_, err := cluster.InitCluster(cfg, true, nil, nil, nil, nil)
+	if err != nil {
+		t.Fatalf("unexpected error: %v", err)
+	}
+}
+
+func TestInitCluster_HealthCheckClientInvalidCAPath(t *testing.T) {
+	t.Parallel()
+	cfg := &kubevip.Config{
+		EnableBGP: true,
+		ControlPlaneHealthCheck: kubevip.HealthCheck{
+			Address: "https://localhost:6443/livez",
+			CAPath:  "/nonexistent/ca.crt",
+		},
+	}
+
+	_, err := cluster.InitCluster(cfg, true, nil, nil, nil, nil)
+	if err == nil {
+		t.Fatal("expected error for invalid CA path")
+	}
+	if !strings.Contains(err.Error(), "reading health check CA cert") {
+		t.Errorf("expected error about reading CA cert, got: %v", err)
+	}
+}
+
+func TestInitCluster_HealthCheckClientInvalidCAContent(t *testing.T) {
+	t.Parallel()
+	caFile := filepath.Join(t.TempDir(), "bad-ca.crt")
+	if err := os.WriteFile(caFile, []byte("not a certificate"), 0o600); err != nil {
+		t.Fatal(err)
+	}
+
+	cfg := &kubevip.Config{
+		EnableBGP: true,
+		ControlPlaneHealthCheck: kubevip.HealthCheck{
+			Address: "https://localhost:6443/livez",
+			CAPath:  caFile,
+		},
+	}
+
+	_, err := cluster.InitCluster(cfg, true, nil, nil, nil, nil)
+	if err == nil {
+		t.Fatal("expected error for invalid CA content")
+	}
+	if !strings.Contains(err.Error(), "contains no valid certificates") {
+		t.Errorf("expected error about invalid certificates, got: %v", err)
+	}
+}
+
+// generateTestCACert creates a self-signed CA certificate in PEM format for testing.
+func generateTestCACert(t *testing.T) []byte {
+	t.Helper()
+
+	key, err := ecdsa.GenerateKey(elliptic.P256(), rand.Reader)
+	if err != nil {
+		t.Fatal(err)
+	}
+
+	template := &x509.Certificate{
+		SerialNumber:          big.NewInt(1),
+		NotBefore:             time.Now(),
+		NotAfter:              time.Now().Add(time.Hour),
+		IsCA:                  true,
+		BasicConstraintsValid: true,
+	}
+
+	certDER, err := x509.CreateCertificate(rand.Reader, template, template, &key.PublicKey, key)
+	if err != nil {
+		t.Fatal(err)
+	}
+
+	return pem.EncodeToMemory(&pem.Block{Type: "CERTIFICATE", Bytes: certDER})
+}

pkg/cluster/service.go

@@ -5,103 +5,136 @@ import (
 	"errors"
 	"fmt"
 	"io/fs"
-	"net"
-	"os"
-	"os/signal"
+	"net/http"
+	"net/url"
+	"strconv"
+	"strings"
 	"sync"
 	"syscall"
+	"time"
 
 	log "log/slog"
 
 	"github.com/kube-vip/kube-vip/pkg/arp"
 	"github.com/kube-vip/kube-vip/pkg/backend"
 	"github.com/kube-vip/kube-vip/pkg/bgp"
+	"github.com/kube-vip/kube-vip/pkg/election"
 	"github.com/kube-vip/kube-vip/pkg/kubevip"
 	"github.com/kube-vip/kube-vip/pkg/loadbalancer"
+	"github.com/kube-vip/kube-vip/pkg/utils"
 	"github.com/kube-vip/kube-vip/pkg/vip"
-	"github.com/vishvananda/netlink"
 	corev1 "k8s.io/api/core/v1"
 	apierrors "k8s.io/apimachinery/pkg/api/errors"
 	metav1 "k8s.io/apimachinery/pkg/apis/meta/v1"
 	"k8s.io/client-go/kubernetes"
 )
 
-func (cluster *Cluster) vipService(ctxArp, ctxDNS context.Context, c *kubevip.Config, sm *Manager, bgpServer *bgp.Server, cancelLeaderElection context.CancelFunc) error {
+// BGPRouteManager allows to manage the routes announced by the BGP server.
+type BGPRouteManager interface {
+	AddHost(ctx context.Context, addr string, object string) error
+	DelHost(ctx context.Context, addr string, object string) error
+}
+
+func (cluster *Cluster) StartVipService(ctx context.Context, c *kubevip.Config, em *election.Manager,
+	bgpServer BGPRouteManager, killFunc func()) error {
+
 	var err error
 
-	// listen for interrupts or the Linux SIGTERM signal and cancel
-	// our context, which the leader election code will observe and
-	// step down
-	signalChan := make(chan os.Signal, 1)
-	// Add Notification for Userland interrupt
-	signal.Notify(signalChan, syscall.SIGINT)
+	var wg sync.WaitGroup
+	defer wg.Wait()
 
-	// Add Notification for SIGTERM (sent from Kubernetes)
-	signal.Notify(signalChan, syscall.SIGTERM)
+	wg.Go(func() {
+		<-ctx.Done()
+		killFunc()
+	})
 
 	loadbalancers := []*loadbalancer.IPVSLoadBalancer{}
 
-	var arpWG sync.WaitGroup
-
 	for i := range cluster.Network {
 		network := cluster.Network[i]
-		if err := network.SetMask(c.VIPSubnet); err != nil {
-			log.Error("failed to set mask", "subnet", c.VIPSubnet, "err", err)
-			panic("")
-		}
+
 		if network.IsDDNS() {
-			if err := cluster.StartDDNS(ctxDNS); err != nil {
-				log.Error(err.Error())
+			if err := cluster.StartDDNS(ctx, cluster.Network[i], c.DHCPBackoffAttempts, &wg); err != nil {
+				log.Error("failed to start DDNS", "err", err)
 			}
 		}
 
+		if err := network.SetMask(c.VIPSubnet); err != nil {
+			killFunc()
+			return fmt.Errorf("failed to set mask for subnet %q: %w", c.VIPSubnet, err)
+		}
+
 		// start the dns updater if address is dns
 		if network.IsDNS() {
 			log.Info("starting the DNS updater", "address", network.DNSName())
 			ipUpdater := vip.NewIPUpdater(network)
-			ipUpdater.Run(ctxDNS)
+			wg.Go(func() {
+				ipUpdater.Run(ctx)
+			})
 		}
 
 		if !c.EnableRoutingTable {
-			if _, err = network.AddIP(false); err != nil {
-				log.Error(err.Error())
+			// Normal VIP addition, use skipDAD=false for normal DAD process
+			if _, err = network.AddIP(false, false); err != nil {
+				log.Error("failed to add IP", "address", network.IP(), "error", err)
 			}
 		}
 
 		if c.EnableBGP {
-			// Lets advertise the VIP over BGP, the host needs to be passed using CIDR notation
-			log.Debug("Attempting to advertise over BGP", "address", network.CIDR())
-			err = bgpServer.AddHost(network.CIDR())
-			if err != nil {
-				log.Error(err.Error())
+			if c.ControlPlaneHealthCheck.Address != "" {
+				// The health check loop owns route advertisement/withdrawal when configured.
+				wg.Go(func() {
+					cluster.bgpHealthCheckLoop(ctx, c, bgpServer, network.CIDR())
+				})
+			} else {
+				// Lets advertise the VIP over BGP, the host needs to be passed using CIDR notation.
+				log.Debug("Attempting to advertise over BGP", "address", network.CIDR())
+				err = bgpServer.AddHost(ctx, network.CIDR(), c.NodeName)
+				if err != nil {
+					log.Error(err.Error())
+				}
 			}
 		}
 
 		if c.EnableLoadBalancer {
-			lb, err := loadbalancer.NewIPVSLB(network.IP(), c.LoadBalancerPort, c.LoadBalancerForwardingMethod, c.BackendHealthCheckInterval, c.Interface, cancelLeaderElection, signalChan)
+			lb, err := loadbalancer.NewIPVSLB(ctx, network, c.LoadBalancerPort, c.LoadBalancerForwardingMethod,
+				c.BackendHealthCheckInterval, c.EgressWithNftables, killFunc, &wg)
 			if err != nil {
-				log.Error("Error creating IPVS LoadBalancer", "err", err)
+				killFunc()
+				return fmt.Errorf("creating IPVS LoadBalancer: %w", err)
 			}
 
-			go func() {
-				err = sm.NodeWatcher(ctxArp, lb, c.Port) //TODO: We're using the ctxARP as the context this will change when rkatz finishes his change
-				if err != nil {
-					log.Error("Error watching node labels", "err", err)
+			wg.Go(func() {
+				for {
+					select {
+					case <-ctx.Done():
+						return
+					default:
+						err = em.NodeWatcher(ctx, lb, c.Port)
+						if err != nil {
+							log.Error("Error watching node labels", "err", err)
+							if utils.IsPanicError(err) {
+								killFunc()
+								return
+							}
+						}
+					}
 				}
-			}()
+			})
 
 			loadbalancers = append(loadbalancers, lb)
 		}
 
 		if c.EnableARP {
-			arpWG.Add(1)
-			go cluster.layer2Update(ctxArp, network, c, &arpWG)
+			wg.Go(func() {
+				cluster.layer2Update(ctx, network, c)
+			})
 		}
 	}
 
 	if c.EnableLoadBalancer {
 		// Shutdown function that will wait on this signal, unless we call it ourselves
-		<-signalChan
+		<-ctx.Done()
 		for _, lb := range loadbalancers {
 			err = lb.RemoveIPVSLB()
 			if err != nil {
@@ -115,64 +148,51 @@ func (cluster *Cluster) vipService(ctxArp, ctxDNS context.Context, c *kubevip.Co
 		backendMapV6 := backend.Map{}
 		// only check localhost
 
-		nodename := ""
-		if c.NodeName != "" {
-			nodename = c.NodeName
+		// An explicitly configured Kubernetes API address (static-pod
+		// deployments point it at the local API server, whose loopback
+		// listener is often the only certificate-valid local endpoint)
+		// takes precedence over the Node object's addresses: the check
+		// answers "is the local API server healthy" for every VIP family,
+		// regardless of the transport family of the override itself.
+		if entry := kubernetesAddrBackendEntry(c.KubernetesAddr, c.Port); entry != nil {
+			log.Info("using configured Kubernetes address for backend health checks", "address", c.KubernetesAddr)
+			backendMapV4[*entry] = false
+			backendMapV6[*entry] = false
 		} else {
-			nodename = os.Getenv("HOSTNAME")
-		}
+			ips := []string{}
+			if c.NodeName != "" {
+				if ips, err = getNodeIPs(ctx, c.NodeName, em.KubernetesClient); err != nil && !apierrors.IsNotFound(err) {
+					log.Error("failed to get IP of control-plane node", "err", err)
+				}
+			}
 
-		ips := []string{}
-		if nodename != "" {
-			if ips, err = getNodeIPs(ctxArp, nodename, sm.KubernetesClient); err != nil && !apierrors.IsNotFound(err) {
-				log.Error("failed to get IP of control-plane nod", "err", err)
+			if len(ips) == 0 {
+				if !utils.IsIPv6(cluster.Network[0].IP()) {
+					ips = append(ips, "127.0.0.1")
+				} else {
+					ips = append(ips, "::1")
+				}
+
+				log.Info("no IP address found for node - will fallback to use localhost address", "addresses", ips)
+			}
+
+			for _, ip := range ips {
+				entry := backend.Entry{Addr: ip, Port: c.Port}
+				if !utils.IsIPv6(ip) {
+					backendMapV4[entry] = false
+				} else {
+					backendMapV6[entry] = false
+				}
 			}
 		}
 
-		if len(ips) == 0 {
-			isV6, err := isV6(cluster.Network[0].IP())
-			if err != nil {
-				return fmt.Errorf("failed to parse IP '%s'", cluster.Network[0].IP())
-			}
-			if !isV6 {
-				ips = append(ips, "127.0.0.1")
-			} else {
-				ips = append(ips, "::1")
-			}
-
-			log.Info("no IP address found for node - will fallback to use localhost address", "addresses", ips)
-		}
-
-		for _, ip := range ips {
-			entry := backend.Entry{Addr: ip, Port: c.Port}
-			ipv6, err := isV6(ip)
-			if err != nil {
-				log.Error("failed to check IP type", "IP", ip, "error", err)
-			}
-			if !ipv6 {
-				backendMapV4[entry] = false
-			} else {
-				backendMapV6[entry] = false
-			}
-		}
-
-		stop := make(chan struct{})
-
-		// will wait for system interrupt and will send stop signal to backend watch
-		go func() {
-			<-signalChan
-			stop <- struct{}{}
-		}()
-
-		backend.Watch(func() {
+		backend.SetKubeConfigPath(c.K8sConfigFile)
+		backend.Watch(ctx, c.BackendHealthCheckInterval, func() {
 			for i := range cluster.Network {
 				network := cluster.Network[i]
 				networkIP := network.IP()
-				isNetworkV6, err := isV6(networkIP)
-				if err != nil {
-					log.Error("failed to check IP type", "IP", networkIP, "error", err)
-					continue
-				}
+				isNetworkV6 := utils.IsIPv6(networkIP)
+				log.Debug("current ip to process", "ip", networkIP)
 
 				backendMap := &backendMapV4
 				if isNetworkV6 {
@@ -180,8 +200,28 @@ func (cluster *Cluster) vipService(ctxArp, ctxDNS context.Context, c *kubevip.Co
 				}
 
 				for entry := range *backendMap {
-					if entry.Check() {
-						_, err = network.AddIP(true)
+					log.Debug("entry.Check() for entry", "entry", entry)
+					var healthy bool
+					if c.ControlPlaneHealthCheck.Address != "" {
+						req, reqErr := http.NewRequestWithContext(ctx, http.MethodGet, c.ControlPlaneHealthCheck.Address, nil)
+						if reqErr != nil {
+							log.Error("create health check request", "err", reqErr)
+						} else if resp, doErr := cluster.healthCheckHTTPClient.Do(req); doErr != nil {
+							log.Error("health check request failed", "url", c.ControlPlaneHealthCheck.Address, "err", doErr)
+						} else {
+							resp.Body.Close()
+							healthy = resp.StatusCode == http.StatusOK
+							if !healthy {
+								log.Warn("health check returned non-200 status", "url", c.ControlPlaneHealthCheck.Address, "status", resp.StatusCode)
+							}
+						}
+					} else {
+						healthy = entry.Check()
+					}
+					if healthy {
+						log.Debug("entry.Check() true")
+						// Normal VIP addition with precheck, use skipDAD=false for normal DAD process
+						_, err = network.AddIP(true, false)
 						if err != nil {
 							log.Error("error adding address", "err", err)
 						}
@@ -189,11 +229,21 @@ func (cluster *Cluster) vipService(ctxArp, ctxDNS context.Context, c *kubevip.Co
 							log.Info("added backend", "ip", network.IP())
 						}
 
-						err = network.AddRoute(true)
+						err = cluster.routeMgr.Add(c.NodeName, network, true, false)
 						if err != nil && !errors.Is(err, fs.ErrExist) && !errors.Is(err, syscall.ESRCH) {
 							log.Warn(err.Error())
 						} else if err == nil && !(*backendMap)[entry] {
-							log.Info("added route", "route", network.PrepareRoute().String())
+							log.Info("added route", "route", network.PrepareRoute())
+						} else if err == nil || errors.Is(err, fs.ErrExist) {
+							// Re-assert the route on every healthy cycle: routing daemons
+							// (e.g. zebra) can miss the single netlink event for the route,
+							// leaving it unadvertised even though it exists in the kernel.
+							// RouteReplace is idempotent and regenerates that event.
+							if replaceErr := network.ReplaceRoute(); replaceErr != nil {
+								log.Warn("re-asserting route", "err", replaceErr)
+							} else {
+								log.Debug("re-asserted route", "route", network.PrepareRoute())
+							}
 						}
 
 						(*backendMap)[entry] = true
@@ -211,35 +261,127 @@ func (cluster *Cluster) vipService(ctxArp, ctxDNS context.Context, c *kubevip.Co
 				}
 
 				if deleteAddress {
-					err = network.DeleteRoute()
-					if err != nil && !errors.Is(err, fs.ErrNotExist) && !errors.Is(err, syscall.ESRCH) {
+					err = cluster.routeMgr.Delete(c.NodeName, network)
+					if err != nil {
 						log.Warn("deleting route", "err", err)
-					} else if err == nil {
-						log.Info("deleted route", "route", network.PrepareRoute().String())
 					}
 
 					deleted, err := network.DeleteIP()
 					if err != nil {
 						log.Error("error deleting IP", "err", err)
-						panic("")
+						killFunc()
+						return
 					}
 					if deleted {
 						log.Info("deleted address", "IP", network.IP(), "interface", network.Interface())
 					}
 				}
 			}
-		}, c.BackendHealthCheckInterval, stop)
+		})
+	}
+
+	if c.EnableBGP {
+		<-ctx.Done()
 	}
 
 	return nil
 }
 
-func isV6(ip string) (bool, error) {
-	ipaddr := net.ParseIP(ip)
-	if ipaddr == nil {
-		return false, fmt.Errorf("failed to parse IP '%s'", ip)
+func (cluster *Cluster) bgpHealthCheckLoop(ctx context.Context, c *kubevip.Config, bgpServer BGPRouteManager, vipCIDR string) {
+	period := time.Duration(c.ControlPlaneHealthCheck.PeriodSeconds) * time.Second
+
+	consecutiveFailures := 0
+	routeAnnounced := false
+	ticker := time.NewTicker(period)
+	defer ticker.Stop()
+
+	log.Info("Starting BGP health check",
+		"address", c.ControlPlaneHealthCheck.Address,
+		"cidr", vipCIDR,
+		"period", period,
+		"timeout", cluster.healthCheckHTTPClient.Timeout,
+		"threshold", c.ControlPlaneHealthCheck.FailureThreshold,
+	)
+
+	for {
+		statusCode := 0
+		var healthErr error
+
+		req, err := http.NewRequestWithContext(ctx, http.MethodGet, c.ControlPlaneHealthCheck.Address, nil)
+		if err != nil {
+			healthErr = err
+		} else {
+			resp, err := cluster.healthCheckHTTPClient.Do(req)
+			if err != nil {
+				healthErr = err
+			} else {
+				defer resp.Body.Close()
+				statusCode = resp.StatusCode
+			}
+		}
+
+		healthy := healthErr == nil && statusCode == http.StatusOK
+
+		if healthy {
+			consecutiveFailures = 0
+			if !routeAnnounced {
+				log.Info("BGP health check passed, announcing route", "cidr", vipCIDR)
+				if err := bgpServer.AddHost(ctx, vipCIDR, c.NodeName); err != nil {
+					log.Error("BGP health check: failed to announce route", "cidr", vipCIDR, "err", err)
+				} else {
+					routeAnnounced = true
+				}
+			}
+		} else {
+			consecutiveFailures++
+			if healthErr != nil {
+				log.Warn("BGP health check failed", "address", c.ControlPlaneHealthCheck.Address, "consecutive", consecutiveFailures, "err", healthErr)
+			} else {
+				log.Warn("BGP health check failed", "address", c.ControlPlaneHealthCheck.Address, "consecutive", consecutiveFailures, "status", statusCode)
+			}
+
+			if consecutiveFailures >= c.ControlPlaneHealthCheck.FailureThreshold && routeAnnounced {
+				log.Warn("BGP health check threshold reached, withdrawing route", "failureThreshold", c.ControlPlaneHealthCheck.FailureThreshold, "cidr", vipCIDR)
+				if err := bgpServer.DelHost(ctx, vipCIDR, c.NodeName); err != nil {
+					log.Error("BGP health check: failed to withdraw route", "cidr", vipCIDR, "err", err)
+				} else {
+					routeAnnounced = false
+				}
+			}
+		}
+
+		select {
+		case <-ctx.Done():
+			if routeAnnounced {
+				if err := bgpServer.DelHost(ctx, vipCIDR, c.NodeName); err != nil {
+					log.Error("BGP health check: failed to withdraw route", "cidr", vipCIDR, "err", err)
+				}
+			}
+			return
+		case <-ticker.C:
+		}
 	}
-	return ipaddr.To4() == nil, nil
+}
+
+// kubernetesAddrBackendEntry converts an explicitly configured Kubernetes
+// API address override (config.KubernetesAddr, e.g. "https://127.0.0.1:6443"
+// on static-pod deployments) into a backend health-check entry. Returns nil
+// when no usable override is configured.
+func kubernetesAddrBackendEntry(kubernetesAddr string, defaultPort uint16) *backend.Entry {
+	if kubernetesAddr == "" {
+		return nil
+	}
+	u, err := url.Parse(kubernetesAddr)
+	if err != nil || u.Hostname() == "" {
+		return nil
+	}
+	port := defaultPort
+	if p := u.Port(); p != "" {
+		if parsed, err := strconv.ParseUint(p, 10, 16); err == nil {
+			port = uint16(parsed)
+		}
+	}
+	return &backend.Entry{Addr: u.Hostname(), Port: port}
 }
 
 func getNodeIPs(ctx context.Context, nodename string, client *kubernetes.Clientset) ([]string, error) {
@@ -257,109 +399,136 @@ func getNodeIPs(ctx context.Context, nodename string, client *kubernetes.Clients
 }
 
 // StartLoadBalancerService will start a VIP instance and leave it for kube-proxy to handle
-func (cluster *Cluster) StartLoadBalancerService(ctx context.Context, c *kubevip.Config, bgp *bgp.Server, name string, CountRouteReferences func(*netlink.Route) int) {
+func (cluster *Cluster) StartLoadBalancerService(ctx context.Context, c *kubevip.Config, bgp *bgp.Server, name string, wg *sync.WaitGroup) error {
 	// use a Go context so we can tell the arp loop code when we
 	// want to step down
 	//nolint
-	ctxArp, cancelArp := context.WithCancel(ctx)
+	lbCtx, lbCancel := context.WithCancel(ctx)
 
-	cluster.stop = make(chan bool, 1)
-	cluster.completed = make(chan bool, 1)
-
-	var arpWG sync.WaitGroup
+	var lbWg sync.WaitGroup
 
 	for i := range cluster.Network {
 		network := cluster.Network[i]
+
+		if network.IsDDNS() {
+			ddnsReady := make(chan struct{})
+			lbWg.Go(func() {
+				// start the DDNS if requested
+				log.Debug("(svcs) start DDNS", "name", network.DNSName())
+				if err := cluster.StartDDNS(lbCtx, cluster.Network[i], c.DHCPBackoffAttempts, &lbWg); err != nil {
+					log.Error("failed to start DDNS", "err", err)
+				}
+
+				close(ddnsReady)
+				<-lbCtx.Done()
+			})
+			<-ddnsReady
+		}
+
+		log.Debug("current ip to process", "ip", network.IP(), "mask", c.VIPSubnet)
 		if err := network.SetMask(c.VIPSubnet); err != nil {
 			log.Error("failed to set mask", "subnet", c.VIPSubnet, "err", err)
-			panic("")
+			lbCancel()
+			return utils.WrapPanicError(err, "failed to set mask for subnet %q", c.VIPSubnet)
 		}
 		_, err := network.DeleteIP()
 		if err != nil {
 			log.Warn("attempted to clean existing VIP", "err", err)
 		}
+		log.Debug("config flags", "enable_routing_table", c.EnableRoutingTable, "enable_leader_election", c.EnableLeaderElection, "enable_services_election", c.EnableServicesElection)
 
 		if c.EnableRoutingTable && (c.EnableLeaderElection || c.EnableServicesElection) {
-			err = network.AddRoute(false)
+			err = cluster.routeMgr.Add(name, network, false, false)
 			if err != nil {
 				log.Warn(err.Error())
+			} else {
+				log.Info("successful add Route")
 			}
-		} else if !c.EnableRoutingTable {
-			if _, err = network.AddIP(false); err != nil {
+		}
+
+		if shouldAddServiceIP(c) {
+			// Normal VIP addition, use skipDAD=false for normal DAD process
+			// Note: When WireGuard is enabled, the VIP is added to the tunnel interface
+			// instead of lo, so we skip adding it here.
+			if _, err = network.AddIP(false, false); err != nil {
 				log.Warn(err.Error())
+			} else {
+				log.Info("successful add IP", "address", network.IP())
 			}
 		}
 
 		if c.EnableARP {
-			arpWG.Add(1)
-			go cluster.layer2Update(ctxArp, network, c, &arpWG)
+			lbWg.Go(func() {
+				cluster.layer2Update(lbCtx, network, c)
+			})
 		}
 
 		if c.EnableBGP && (c.EnableLeaderElection || c.EnableServicesElection) {
 			// Lets advertise the VIP over BGP, the host needs to be passed using CIDR notation
 			log.Debug("(svcs) attempting to advertise over BGP", "address", network.CIDR())
-			err = bgp.AddHost(network.CIDR())
+			err = bgp.AddHost(lbCtx, network.CIDR(), name)
 			if err != nil {
 				log.Error(err.Error())
 			}
 		}
 	}
 
-	go func() {
-		<-cluster.stop
-		// Stop the Arp context if it is running
-		cancelArp()
+	wg.Go(func() {
+		for i := range cluster.Network {
+			network := cluster.Network[i]
 
-		arpWG.Wait() // wait for all cluster ARP/NDP to be finished
+			// start the dns updater if address is dns
+			if network.IsDNS() {
+				log.Info("(svcs) starting the DNS updater", "address", network.DNSName(), "ip", network.IP())
+				ipUpdater := vip.NewIPUpdater(network)
+				wg.Go(func() {
+					ipUpdater.Run(lbCtx)
+				})
+			}
+		}
+
+		select {
+		case <-cluster.stop:
+		case <-ctx.Done():
+		}
+
+		// Stop the loadbalancer context if it is running
+		lbCancel()
+
+		lbWg.Wait() // wait for all cluster ARP/NDP to be finished
 
 		log.Info("[LOADBALANCER] Stopping load balancers", "name", name)
 
 		if c.EnableRoutingTable {
 			for i := range cluster.Network {
-				// chek if route is not  referenced by another service
-				r := cluster.Network[i].PrepareRoute()
-				if CountRouteReferences(r) < 1 {
-					log.Info("[VIP] Deleting Route for VIP", "IP", cluster.Network[i].IP())
-					if err := cluster.Network[i].DeleteRoute(); err != nil {
-						log.Warn(err.Error())
-					}
+				if err := cluster.routeMgr.Delete(name, cluster.Network[i]); err != nil {
+					log.Warn(err.Error())
 				}
 			}
 
-			close(cluster.completed)
 			return
 		}
-		for i := range cluster.Network {
-			if c.EnableARP && cluster.arpMgr.Count(cluster.Network[i].ARPName()) > 0 {
-				continue
-			}
-			log.Info("[VIP] Deleting VIP", "ip", cluster.Network[i].IP())
-			deleted, err := cluster.Network[i].DeleteIP()
-			if err != nil {
-				log.Warn(err.Error())
-			}
-			if deleted {
-				log.Info("deleted address", "IP", cluster.Network[i].IP(), "interface", cluster.Network[i].Interface())
-			}
-		}
 
-		close(cluster.completed)
-	}()
+		cluster.cleanupVIPs(c)
+	})
+
+	return nil
+}
+
+func shouldAddServiceIP(c *kubevip.Config) bool {
+	return !c.EnableRoutingTable && (!c.EnableBGP || c.BGPAttachIPToInterface) && !c.EnableWireguard
 }
 
 // Layer2Update, handles the creation of the
-func (cluster *Cluster) layer2Update(ctx context.Context, network vip.Network, c *kubevip.Config, arpWG *sync.WaitGroup) {
-	defer arpWG.Done()
-	log.Info("layer 2 broadcaster starting")
+func (cluster *Cluster) layer2Update(ctx context.Context, network vip.Network, c *kubevip.Config) {
 	var ndp *vip.NdpResponder
 	var err error
 	ipString := network.IP()
-	if vip.IsIPv6(ipString) {
+	if utils.IsIPv6(ipString) {
 		if network.IPisLinkLocal() {
 			log.Error("layer2 is link-local can't use NDP", "address", ipString)
-
 		} else {
-			ndp, err = vip.NewNDPResponder(network.Interface())
+			ndp, err = waitNDPResponder(ctx, network.Interface())
 			if err != nil {
 				log.Error("failed to create new NDP Responder", "error", err)
 			} else {
@@ -370,6 +539,7 @@ func (cluster *Cluster) layer2Update(ctx context.Context, network vip.Network, c
 		}
 	}
 
+	log.Info("layer 2 broadcaster starting", "IP", network.IP(), "device", network.Interface())
 	log.Debug("layer 2 update", "ip", ipString, "interface", network.Interface(), "ms", c.ArpBroadcastRate)
 
 	arpInstance := arp.NewInstance(network, ndp)
@@ -377,5 +547,32 @@ func (cluster *Cluster) layer2Update(ctx context.Context, network vip.Network, c
 
 	<-ctx.Done() // if cancel() execute
 	log.Debug("ending layer 2 update", "ip", ipString, "interface", network.Interface(), "ms", c.ArpBroadcastRate)
-	cluster.arpMgr.Remove(arpInstance)
+	cluster.arpMgr.RemoveOnLeadershipLoss(arpInstance)
+}
+
+func waitNDPResponder(ctx context.Context, ifaceName string) (*vip.NdpResponder, error) {
+	ndp, err := vip.NewNDPResponder(ifaceName)
+	if err != nil && strings.Contains(err.Error(), "no such device") {
+		log.Warn("unable to create NDP responder at first try", "interface", ifaceName, "err", err)
+		ndpCreateCtx, cancel := context.WithTimeout(ctx, 2*time.Minute)
+		defer cancel()
+		ticker := time.NewTicker(time.Second)
+
+		for {
+			select {
+			case <-ndpCreateCtx.Done():
+				return nil, fmt.Errorf("failed to create NDP responder for interface %q: %w", ifaceName, ndpCreateCtx.Err())
+			case <-ticker.C:
+				ndp, err = vip.NewNDPResponder(ifaceName)
+				if err != nil {
+					log.Warn("unable to create NDP responder on retry", "interface", ifaceName, "err", err)
+				} else {
+					return ndp, nil
+				}
+			}
+		}
+	} else if err != nil {
+		return nil, fmt.Errorf("unable to create NDP responder for interface %q: %w", ifaceName, err)
+	}
+	return ndp, nil
 }

pkg/cluster/service_config_test.go

@@ -0,0 +1,55 @@
+package cluster
+
+import (
+	"testing"
+
+	"github.com/kube-vip/kube-vip/pkg/kubevip"
+)
+
+func TestShouldAddServiceIP(t *testing.T) {
+	tests := []struct {
+		name   string
+		config *kubevip.Config
+		want   bool
+	}{
+		{
+			name:   "BGP default does not attach IP",
+			config: &kubevip.Config{EnableBGP: true},
+			want:   false,
+		},
+		{
+			name: "BGP opt-in attaches IP",
+			config: &kubevip.Config{
+				EnableBGP:              true,
+				BGPAttachIPToInterface: true,
+			},
+			want: true,
+		},
+		{
+			name: "routing table takes precedence",
+			config: &kubevip.Config{
+				EnableBGP:              true,
+				BGPAttachIPToInterface: true,
+				EnableRoutingTable:     true,
+			},
+			want: false,
+		},
+		{
+			name: "WireGuard takes precedence",
+			config: &kubevip.Config{
+				EnableBGP:              true,
+				BGPAttachIPToInterface: true,
+				EnableWireguard:        true,
+			},
+			want: false,
+		},
+	}
+
+	for _, tt := range tests {
+		t.Run(tt.name, func(t *testing.T) {
+			if got := shouldAddServiceIP(tt.config); got != tt.want {
+				t.Fatalf("shouldAddServiceIP() = %t, want %t", got, tt.want)
+			}
+		})
+	}
+}

pkg/cluster/service_internal_test.go

@@ -0,0 +1,61 @@
+package cluster
+
+import (
+	"testing"
+)
+
+func TestKubernetesAddrBackendEntry(t *testing.T) {
+	cases := []struct {
+		name     string
+		addr     string
+		port     uint16
+		wantAddr string
+		wantPort uint16
+		wantNil  bool
+	}{
+		{
+			name:     "explicit v4 loopback with port",
+			addr:     "https://127.0.0.1:6443",
+			port:     9999,
+			wantAddr: "127.0.0.1",
+			wantPort: 6443,
+		},
+		{
+			name:     "hostname without port falls back to config port",
+			addr:     "https://localhost",
+			port:     6443,
+			wantAddr: "localhost",
+			wantPort: 6443,
+		},
+		{
+			name:    "empty override",
+			addr:    "",
+			port:    6443,
+			wantNil: true,
+		},
+		{
+			name:    "garbage override",
+			addr:    "://not-a-url",
+			port:    6443,
+			wantNil: true,
+		},
+	}
+
+	for _, tc := range cases {
+		t.Run(tc.name, func(t *testing.T) {
+			entry := kubernetesAddrBackendEntry(tc.addr, tc.port)
+			if tc.wantNil {
+				if entry != nil {
+					t.Fatalf("expected nil entry, got %+v", entry)
+				}
+				return
+			}
+			if entry == nil {
+				t.Fatal("expected an entry, got nil")
+			}
+			if entry.Addr != tc.wantAddr || entry.Port != tc.wantPort {
+				t.Fatalf("got %+v, want addr %q port %d", entry, tc.wantAddr, tc.wantPort)
+			}
+		})
+	}
+}

pkg/cluster/service_test.go

@@ -0,0 +1,430 @@
+package cluster_test
+
+import (
+	"context"
+	"encoding/pem"
+	"net/http"
+	"net/http/httptest"
+	"os"
+	"path/filepath"
+	"sync"
+	"sync/atomic"
+	"testing"
+	"time"
+
+	"github.com/kube-vip/kube-vip/pkg/cluster"
+	"github.com/kube-vip/kube-vip/pkg/kubevip"
+	"github.com/kube-vip/kube-vip/pkg/route"
+	"github.com/kube-vip/kube-vip/pkg/vip"
+	"github.com/vishvananda/netlink"
+	corev1 "k8s.io/api/core/v1"
+)
+
+const testCIDR = "10.0.0.34/32"
+
+func TestBGPHealthCheckLoop_AnnouncesOnHealthy(t *testing.T) {
+	t.Parallel()
+	healthcheck := newTestHealthServer(t, http.StatusOK)
+	t.Cleanup(healthcheck.server.Close)
+
+	bgpManager := newMockBGPRouteManager()
+	startVipService(t, newTestConfig(healthcheck.server.URL, healthcheck.caPath), bgpManager)
+
+	expectEventually(t, func() bool { return bgpManager.isAnnounced() },
+		"route should be announced")
+}
+
+func TestBGPHealthCheckLoop_NoAnnouncementUntilHealthy(t *testing.T) {
+	t.Parallel()
+	healthcheck := newTestHealthServer(t, http.StatusInternalServerError)
+	t.Cleanup(healthcheck.server.Close)
+
+	bgpManager := newMockBGPRouteManager()
+	startVipService(t, newTestConfig(healthcheck.server.URL, healthcheck.caPath), bgpManager)
+
+	expectConsistently(t, func() bool { return !bgpManager.isAnnounced() },
+		2*time.Second, "route should not be announced while unhealthy")
+
+	healthcheck.setStatus(http.StatusOK)
+	expectEventually(t, func() bool { return bgpManager.isAnnounced() },
+		"route should be announced after recovery")
+}
+
+func TestBGPHealthCheckLoop_WithdrawsAfterThreshold(t *testing.T) {
+	t.Parallel()
+	healthcheck := newTestHealthServer(t, http.StatusOK)
+	t.Cleanup(healthcheck.server.Close)
+
+	bgpManager := newMockBGPRouteManager()
+	cfg := newTestConfig(healthcheck.server.URL, healthcheck.caPath)
+	cfg.ControlPlaneHealthCheck.FailureThreshold = 3
+	startVipService(t, cfg, bgpManager)
+
+	expectEventually(t, func() bool { return bgpManager.isAnnounced() },
+		"route should be announced")
+
+	healthcheck.setStatus(http.StatusServiceUnavailable)
+
+	expectConsistently(t, func() bool { return bgpManager.isAnnounced() },
+		1500*time.Millisecond, "route should stay announced before threshold is reached")
+
+	expectEventually(t, func() bool { return !bgpManager.isAnnounced() },
+		"route should be withdrawn after threshold")
+}
+
+func TestBGPHealthCheckLoop_ReAnnouncesOnRecovery(t *testing.T) {
+	t.Parallel()
+	healthcheck := newTestHealthServer(t, http.StatusOK)
+	t.Cleanup(healthcheck.server.Close)
+
+	bgpManager := newMockBGPRouteManager()
+	cfg := newTestConfig(healthcheck.server.URL, healthcheck.caPath)
+	cfg.ControlPlaneHealthCheck.FailureThreshold = 1
+	startVipService(t, cfg, bgpManager)
+
+	expectEventually(t, func() bool { return bgpManager.isAnnounced() },
+		"route should be announced")
+
+	healthcheck.setStatus(http.StatusServiceUnavailable)
+	expectEventually(t, func() bool { return !bgpManager.isAnnounced() },
+		"route should be withdrawn")
+
+	healthcheck.setStatus(http.StatusOK)
+	expectEventually(t, func() bool { return bgpManager.isAnnounced() },
+		"route should be re-announced")
+}
+
+func TestBGPHealthCheckLoop_StopsOnContextCancel(t *testing.T) {
+	t.Parallel()
+	healthcheck := newTestHealthServer(t, http.StatusOK)
+	t.Cleanup(healthcheck.server.Close)
+
+	bgpManager := newMockBGPRouteManager()
+	cancelContext, vipServiceDone := startVipService(t, newTestConfig(healthcheck.server.URL, healthcheck.caPath), bgpManager)
+
+	expectEventually(t, func() bool { return bgpManager.isAnnounced() },
+		"route should be announced")
+
+	cancelContext()
+
+	select {
+	case <-vipServiceDone:
+	case <-time.After(5 * time.Second):
+		t.Fatal("vipService did not stop after context cancellation")
+	}
+}
+
+func TestBGPHealthCheckLoop_RetriesAddHostOnFailure(t *testing.T) {
+	t.Parallel()
+	healthcheck := newTestHealthServer(t, http.StatusOK)
+	t.Cleanup(healthcheck.server.Close)
+
+	bgpManager := newMockBGPRouteManager()
+	bgpManager.setAddErr(errTestAddHost)
+	startVipService(t, newTestConfig(healthcheck.server.URL, healthcheck.caPath), bgpManager)
+
+	expectConsistently(t, func() bool { return !bgpManager.isAnnounced() },
+		2*time.Second, "route should not be announced while AddHost errors")
+
+	bgpManager.setAddErr(nil)
+	expectEventually(t, func() bool { return bgpManager.isAnnounced() },
+		"route should be announced after clearing AddHost error")
+}
+
+func TestBGPHealthCheckLoop_RetriesDelHostOnFailure(t *testing.T) {
+	t.Parallel()
+	healthcheck := newTestHealthServer(t, http.StatusOK)
+	t.Cleanup(healthcheck.server.Close)
+
+	bgpManager := newMockBGPRouteManager()
+	cfg := newTestConfig(healthcheck.server.URL, healthcheck.caPath)
+	cfg.ControlPlaneHealthCheck.FailureThreshold = 1
+	startVipService(t, cfg, bgpManager)
+
+	expectEventually(t, func() bool { return bgpManager.isAnnounced() },
+		"route should be announced")
+
+	bgpManager.setDelErr(errTestDelHost)
+	healthcheck.setStatus(http.StatusServiceUnavailable)
+
+	expectConsistently(t, func() bool { return bgpManager.isAnnounced() },
+		1500*time.Millisecond, "route should stay announced while DelHost errors")
+
+	bgpManager.setDelErr(nil)
+	expectEventually(t, func() bool { return !bgpManager.isAnnounced() },
+		"route should be withdrawn after clearing DelHost error")
+}
+
+func TestRoutingTableHealthCheck_AddsVIPWhenHealthy(t *testing.T) {
+	t.Parallel()
+	healthcheck := newTestHealthServer(t, http.StatusOK)
+	t.Cleanup(healthcheck.server.Close)
+
+	network := &mockNetwork{ip: "10.0.0.1", cidr: testCIDR}
+	startRoutingTableVipService(t, newRoutingTableConfig(healthcheck.server.URL, healthcheck.caPath), network)
+
+	expectEventually(t, network.isPresent,
+		"VIP should be added while health check is healthy")
+}
+
+func TestRoutingTableHealthCheck_RemovesVIPWhenUnhealthy(t *testing.T) {
+	t.Parallel()
+	healthcheck := newTestHealthServer(t, http.StatusOK)
+	t.Cleanup(healthcheck.server.Close)
+
+	network := &mockNetwork{ip: "10.0.0.1", cidr: testCIDR}
+	startRoutingTableVipService(t, newRoutingTableConfig(healthcheck.server.URL, healthcheck.caPath), network)
+
+	expectEventually(t, network.isPresent,
+		"VIP should be added while health check is healthy")
+
+	healthcheck.setStatus(http.StatusServiceUnavailable)
+	expectEventually(t, func() bool { return !network.isPresent() },
+		"VIP should be removed once health check becomes unhealthy")
+}
+
+var (
+	errTestAddHost = &testError{msg: "mock AddHost error"}
+	errTestDelHost = &testError{msg: "mock DelHost error"}
+)
+
+type testError struct{ msg string }
+
+func (e *testError) Error() string { return e.msg }
+
+// startVipService launches vipService in a goroutine with a mock network and
+// registers a cleanup to cancel the context and wait for it to finish.
+// Uses InitCluster so the real code parses certs for the BGP health check client.
+func startVipService(t *testing.T, cfg *kubevip.Config, bgpManager *mockBGPRouteManager) (context.CancelFunc, <-chan struct{}) {
+	t.Helper()
+
+	c, err := cluster.InitCluster(cfg, true, nil, nil, nil, nil)
+	if err != nil {
+		t.Fatalf("InitCluster: %v", err)
+	}
+	c.Network = []vip.Network{&mockNetwork{ip: "10.0.0.1", cidr: testCIDR}}
+
+	ctx, cancel := context.WithCancel(context.Background())
+	done := make(chan struct{})
+
+	go func() {
+		_ = c.StartVipService(ctx, cfg, nil, bgpManager, func() {})
+		close(done)
+	}()
+
+	t.Cleanup(func() {
+		cancel()
+		<-done
+	})
+
+	return cancel, done
+}
+
+// startRoutingTableVipService launches vipService in routing-table mode with a
+// mock network and a real route.Manager (which only drives the mock network's
+// route methods, so no netlink calls happen). Registers cleanup to stop it.
+func startRoutingTableVipService(t *testing.T, cfg *kubevip.Config, network *mockNetwork) {
+	t.Helper()
+
+	c, err := cluster.InitCluster(cfg, true, nil, nil, route.NewManager(), nil)
+	if err != nil {
+		t.Fatalf("InitCluster: %v", err)
+	}
+	c.Network = []vip.Network{network}
+
+	ctx, cancel := context.WithCancel(context.Background())
+	done := make(chan struct{})
+
+	go func() {
+		_ = c.StartVipService(ctx, cfg, nil, nil, func() {})
+		close(done)
+	}()
+
+	t.Cleanup(func() {
+		cancel()
+		<-done
+	})
+}
+
+func newRoutingTableConfig(url, caPath string) *kubevip.Config {
+	cfg := newTestConfig(url, caPath)
+	cfg.EnableBGP = false
+	cfg.EnableRoutingTable = true
+	cfg.BackendHealthCheckInterval = 1
+	return cfg
+}
+
+func newTestConfig(url, caPath string) *kubevip.Config {
+	return &kubevip.Config{
+		EnableBGP: true,
+		ControlPlaneHealthCheck: kubevip.HealthCheck{
+			Address:          url,
+			CAPath:           caPath,
+			PeriodSeconds:    1,
+			TimeoutSeconds:   2,
+			FailureThreshold: 1,
+		},
+	}
+}
+
+// mockBGPRouteManager tracks announced addresses as a set.
+// AddHost adds, DelHost removes. Errors prevent state changes.
+type mockBGPRouteManager struct {
+	mu        sync.Mutex
+	announced map[string]bool
+	addErr    error
+	delErr    error
+}
+
+func newMockBGPRouteManager() *mockBGPRouteManager {
+	return &mockBGPRouteManager{announced: make(map[string]bool)}
+}
+
+func (m *mockBGPRouteManager) AddHost(_ context.Context, addr string, _ string) error {
+	m.mu.Lock()
+	defer m.mu.Unlock()
+	if m.addErr != nil {
+		return m.addErr
+	}
+	m.announced[addr] = true
+	return nil
+}
+
+func (m *mockBGPRouteManager) DelHost(_ context.Context, addr string, _ string) error {
+	m.mu.Lock()
+	defer m.mu.Unlock()
+	if m.delErr != nil {
+		return m.delErr
+	}
+	delete(m.announced, addr)
+	return nil
+}
+
+func (m *mockBGPRouteManager) isAnnounced() bool {
+	m.mu.Lock()
+	defer m.mu.Unlock()
+	return m.announced[testCIDR]
+}
+
+func (m *mockBGPRouteManager) setAddErr(err error) {
+	m.mu.Lock()
+	m.addErr = err
+	m.mu.Unlock()
+}
+
+func (m *mockBGPRouteManager) setDelErr(err error) {
+	m.mu.Lock()
+	m.delErr = err
+	m.mu.Unlock()
+}
+
+// mockNetwork implements vip.Network with no-op operations.
+type mockNetwork struct {
+	ip   string
+	cidr string
+
+	mu      sync.Mutex
+	present bool
+}
+
+func (m *mockNetwork) AddIP(bool, bool, ...int) (bool, error) {
+	m.mu.Lock()
+	defer m.mu.Unlock()
+	m.present = true
+	return true, nil
+}
+func (m *mockNetwork) DeleteIP() (bool, error) {
+	m.mu.Lock()
+	defer m.mu.Unlock()
+	m.present = false
+	return false, nil
+}
+func (m *mockNetwork) isPresent() bool {
+	m.mu.Lock()
+	defer m.mu.Unlock()
+	return m.present
+}
+func (m *mockNetwork) AddRoute(bool) (bool, error)     { return false, nil }
+func (m *mockNetwork) ReplaceRoute() error             { return nil }
+func (m *mockNetwork) DeleteRoute() error              { return nil }
+func (m *mockNetwork) UpdateRoutes() (bool, error)     { return false, nil }
+func (m *mockNetwork) IsSet() (*netlink.Addr, error)   { return nil, nil }
+func (m *mockNetwork) IP() string                      { return m.ip }
+func (m *mockNetwork) CIDR() string                    { return m.cidr }
+func (m *mockNetwork) IPisLinkLocal() bool             { return false }
+func (m *mockNetwork) PrepareRoute() *netlink.Route    { return nil }
+func (m *mockNetwork) RouteHash() string               { return "" }
+func (m *mockNetwork) SetIP(string) error              { return nil }
+func (m *mockNetwork) SetServicePorts(*corev1.Service) {}
+func (m *mockNetwork) Interface() string               { return "eth0" }
+func (m *mockNetwork) IsDADFAIL() bool                 { return false }
+func (m *mockNetwork) IsDNS() bool                     { return false }
+func (m *mockNetwork) IsDDNS() bool                    { return false }
+func (m *mockNetwork) DDNSHostName() string            { return "" }
+func (m *mockNetwork) DNSName() string                 { return "" }
+func (m *mockNetwork) SetMask(string) error            { return nil }
+func (m *mockNetwork) SetHasEndpoints(bool)            {}
+func (m *mockNetwork) HasEndpoints() bool              { return false }
+func (m *mockNetwork) ARPName() string                 { return "" }
+func (m *mockNetwork) GetPossibleSubnets() string      { return "" }
+func (m *mockNetwork) DHCPFamily() string              { return "" }
+func (m *mockNetwork) IPVSMark() uint32                { return 0 }
+
+// testHealthServer wraps an HTTPS httptest.Server with an atomic status code.
+// caPath is the path to the server's CA cert for client verification.
+type testHealthServer struct {
+	server     *httptest.Server
+	statusCode atomic.Int64
+	caPath     string
+}
+
+func newTestHealthServer(t *testing.T, status int) *testHealthServer {
+	t.Helper()
+	healthcheck := &testHealthServer{}
+	healthcheck.statusCode.Store(int64(status))
+	healthcheck.server = httptest.NewTLSServer(http.HandlerFunc(func(w http.ResponseWriter, _ *http.Request) {
+		w.WriteHeader(int(healthcheck.statusCode.Load()))
+	}))
+
+	cert := healthcheck.server.Certificate()
+	if cert == nil {
+		t.Fatal("TLS server has no certificate")
+	}
+	caPEM := pem.EncodeToMemory(&pem.Block{Type: "CERTIFICATE", Bytes: cert.Raw})
+	caFile := filepath.Join(t.TempDir(), "ca.crt")
+	if err := os.WriteFile(caFile, caPEM, 0o600); err != nil {
+		t.Fatal(err)
+	}
+	healthcheck.caPath = caFile
+	return healthcheck
+}
+
+func (ths *testHealthServer) setStatus(code int) {
+	ths.statusCode.Store(int64(code))
+}
+
+// expectConsistently continuously checks that condition remains true for the given duration.
+// Fails immediately if the condition becomes false at any point.
+func expectConsistently(t *testing.T, condition func() bool, duration time.Duration, msg string) {
+	t.Helper()
+	deadline := time.Now().Add(duration)
+	for time.Now().Before(deadline) {
+		if !condition() {
+			t.Fatalf("condition violated: %s", msg)
+		}
+		time.Sleep(50 * time.Millisecond)
+	}
+}
+
+// expectEventually polls condition until it returns true or 5s timeout is reached.
+func expectEventually(t *testing.T, condition func() bool, msg string) {
+	t.Helper()
+	deadline := time.Now().Add(5 * time.Second)
+	for time.Now().Before(deadline) {
+		if condition() {
+			return
+		}
+		time.Sleep(50 * time.Millisecond)
+	}
+	t.Fatalf("timeout: %s", msg)
+}

pkg/cluster/singleNode.go

@@ -1,84 +0,0 @@
-package cluster
-
-import (
-	"context"
-
-	log "log/slog"
-
-	"github.com/kube-vip/kube-vip/pkg/bgp"
-	"github.com/kube-vip/kube-vip/pkg/kubevip"
-	"github.com/kube-vip/kube-vip/pkg/vip"
-)
-
-// StartSingleNode will start a single node cluster
-func (cluster *Cluster) StartSingleNode(c *kubevip.Config, disableVIP bool) error {
-	// Start kube-vip as a single node server
-
-	// TODO - Split all this code out as a separate function
-	log.Info("Starting kube-vip as a single node cluster")
-
-	log.Info("This node is assuming leadership of the cluster")
-
-	cluster.stop = make(chan bool, 1)
-	cluster.completed = make(chan bool, 1)
-
-	for i := range cluster.Network {
-		if !disableVIP {
-			deleted, err := cluster.Network[i].DeleteIP()
-			if err != nil {
-				log.Warn("Attempted to clean existing VIP", "err", err)
-			}
-			if deleted {
-				log.Info("deleted address", "IP", cluster.Network[i].IP(), "interface", cluster.Network[i].Interface())
-			}
-
-			_, err = cluster.Network[i].AddIP(false)
-			if err != nil {
-				log.Warn(err.Error())
-			}
-
-		}
-
-		if c.EnableARP {
-			// Gratuitous ARP, will broadcast to new MAC <-> IP
-			err := vip.ARPSendGratuitous(cluster.Network[i].IP(), c.Interface)
-			if err != nil {
-				log.Warn(err.Error())
-			}
-		}
-	}
-
-	go func() {
-		<-cluster.stop
-
-		if !disableVIP {
-			for i := range cluster.Network {
-				log.Info("[VIP] Releasing the VIP", "address", cluster.Network[i].IP())
-				deleted, err := cluster.Network[i].DeleteIP()
-				if err != nil {
-					log.Warn(err.Error())
-				}
-				if deleted {
-					log.Info("deleted address", "IP", cluster.Network[i].IP(), "interface", cluster.Network[i].Interface())
-				}
-			}
-		}
-		close(cluster.completed)
-	}()
-	log.Info("Started Load Balancer and Virtual IP")
-	return nil
-}
-
-func (cluster *Cluster) StartVipService(c *kubevip.Config, sm *Manager, bgp *bgp.Server) error {
-	// use a Go context so we can tell the arp loop code when we
-	// want to step down
-	ctxArp, cancelArp := context.WithCancel(context.Background())
-	defer cancelArp()
-
-	// use a Go context so we can tell the dns loop code when we
-	// want to step down
-	ctxDNS, cancelDNS := context.WithCancel(context.Background())
-	defer cancelDNS()
-
-	return cluster.vipService(ctxArp, ctxDNS, c, sm, bgp, nil)
-}

pkg/debouncer/debouncer.go

@@ -0,0 +1,252 @@
+package debouncer
+
+import (
+	"context"
+	"fmt"
+	log "log/slog"
+	"sync"
+	"sync/atomic"
+	"time"
+
+	v1 "k8s.io/api/core/v1"
+	discoveryv1 "k8s.io/api/discovery/v1"
+	"k8s.io/apimachinery/pkg/watch"
+)
+
+const (
+	DefaultTime = "0s"
+	minimalTime = time.Millisecond * 200
+)
+
+type debouncer struct {
+	input    <-chan watch.Event
+	output   chan watch.Event
+	stopChan chan any
+	stopOnce sync.Once
+	// events holds event per namespace
+	namespaces   sync.Map
+	debounceTime time.Duration
+}
+
+type ns struct {
+	sync.Map
+	cnt atomic.Int64
+}
+
+func (n *ns) get(name string) (*object, bool) {
+	value, exists := n.Load(name)
+	if !exists {
+		return nil, false
+	}
+	i, ok := value.(*object)
+	if !ok {
+		return nil, false
+	}
+	return i, true
+}
+
+func (n *ns) add(name string, output chan<- watch.Event) *object {
+	i := newObject(output)
+	n.Store(name, i)
+	n.cnt.Add(1)
+	return i
+}
+
+func (n *ns) del(name string) {
+	if _, exists := n.Load(name); exists {
+		n.Delete(name)
+		n.cnt.Add(-1)
+	}
+}
+
+func New(input <-chan watch.Event, debounceTime string) (*debouncer, error) {
+	dt, err := time.ParseDuration(debounceTime)
+	if err != nil {
+		// debouncer was configured with invalid unparsable value, return error
+		return nil, fmt.Errorf("failed to parse debounce time configuration: %w", err)
+	}
+	if dt < minimalTime {
+		if dt > 0 {
+			log.Warn("configured debounce time is less than the minimal threshold of 200ms, debouncer will remain disabled", "config value", dt.String())
+		}
+		return nil, nil
+	}
+	return &debouncer{
+		input:        input,
+		output:       make(chan watch.Event),
+		stopChan:     make(chan any),
+		debounceTime: dt,
+	}, nil
+}
+
+func (d *debouncer) Start(ctx context.Context) error {
+	wg := sync.WaitGroup{}
+	debouncerCtx, cancel := context.WithCancel(ctx)
+	defer func() {
+		cancel()
+		wg.Wait()
+		close(d.output)
+	}()
+
+	for {
+		select {
+		case <-debouncerCtx.Done():
+			// return if debouncer context was cancelled
+			return nil
+		case <-d.stopChan:
+			// return if Stop() was called
+			return nil
+		case tmp := <-d.input:
+			// event has no type, probably error
+			if tmp.Type == "" {
+				return fmt.Errorf("get undefined object (input channel probably closed)")
+			}
+
+			var namespace, name string
+
+			// type switch event object
+			switch v := tmp.Object.(type) {
+			case *discoveryv1.EndpointSlice:
+				namespace = v.Namespace
+				name = v.Name
+			case *v1.Endpoints: //nolint:staticcheck
+				namespace = v.Namespace
+				name = v.Name
+			case *v1.Service:
+				namespace = v.Namespace
+				name = v.Name
+			default:
+				return fmt.Errorf("objects of type %T are not supported", v)
+			}
+
+			eventNs, exists := d.getNs(namespace)
+			if !exists {
+				// if not, create new map for the namespace
+				eventNs = d.addNs(namespace)
+			}
+
+			// check if the object was previously reconciled
+			eventObject, exists := eventNs.get(name)
+
+			// if not and the event is not of type 'Deleted', create new object
+			if !exists && tmp.Type != watch.Deleted {
+				eventObject = eventNs.add(name, d.output)
+
+				wg.Go(func() {
+					// start deboucing events for this object
+					eventObject.start(debouncerCtx, d.debounceTime)
+					// if debouncer for the object ended - e.g. object was deleted - clean the map of objects
+					eventObject = nil
+					eventNs.del(name)
+					// if namespace is empty, delete the namespace map
+					if eventNs.cnt.Load() == 0 {
+						d.delNs(namespace)
+					}
+				})
+			}
+
+			if eventObject != nil {
+				// pass the watch event to the debouncer object
+				eventObject.input <- tmp
+			}
+		}
+	}
+}
+
+func (d *debouncer) Stop() {
+	d.stopOnce.Do(func() {
+		close(d.stopChan)
+	})
+}
+
+func (d *debouncer) Output() chan watch.Event {
+	return d.output
+}
+
+func (d *debouncer) getNs(namespace string) (*ns, bool) {
+	value, exists := d.namespaces.Load(namespace)
+	if !exists {
+		return nil, false
+	}
+	n, ok := value.(*ns)
+	if !ok {
+		return nil, false
+	}
+	return n, true
+}
+
+func (d *debouncer) addNs(namespace string) *ns {
+	n := ns{}
+	d.namespaces.Store(namespace, &n)
+	return &n
+}
+
+func (d *debouncer) delNs(namespace string) {
+	d.namespaces.Delete(namespace)
+}
+
+type object struct {
+	input    chan watch.Event
+	output   chan<- watch.Event
+	stopChan chan any
+	stopOnce sync.Once
+}
+
+func newObject(output chan<- watch.Event) *object {
+	return &object{
+		input:    make(chan watch.Event),
+		output:   output,
+		stopChan: make(chan any),
+	}
+}
+
+func (o *object) start(ctx context.Context, debounceTime time.Duration) {
+	t := time.NewTicker(debounceTime)
+
+	var last *watch.Event
+
+	defer func() {
+		if last != nil {
+			o.output <- *last
+			last = nil
+		}
+	}()
+
+	for {
+		select {
+		case <-ctx.Done():
+			// if context is done, return
+			return
+		case <-o.stopChan:
+			// return if Stop() was called
+			return
+		case tmp := <-o.input:
+			// if last event is known, but an event of another type arrived,
+			// send out the previous event
+			if last != nil && last.Type != tmp.Type {
+				o.output <- *last
+			}
+			// save current event as the last event
+			last = &tmp
+			// reset the ticker to wait for more events
+			t.Reset(debounceTime)
+		case <-t.C:
+			if last != nil {
+				// on tick, if we have an event, send it out
+				o.output <- *last
+				// if the event is of type 'Deleted', stop the debouncer for the object
+				if last.Type == watch.Deleted {
+					o.stop()
+				}
+				// reset last known event, so it won't be send out twice
+				last = nil
+			}
+		}
+	}
+}
+
+func (o *object) stop() {
+	o.stopOnce.Do(func() {
+		close(o.stopChan)
+	})
+}

pkg/debouncer/debouncer_test.go

@@ -0,0 +1,394 @@
+package debouncer
+
+import (
+	"context"
+	"fmt"
+	"strconv"
+	"sync"
+	"testing"
+	"time"
+
+	v1 "k8s.io/api/core/v1"
+	discoveryv1 "k8s.io/api/discovery/v1"
+	metav1 "k8s.io/apimachinery/pkg/apis/meta/v1"
+	"k8s.io/apimachinery/pkg/watch"
+)
+
+func TestTimeSetting(t *testing.T) {
+	tcs := []struct {
+		name       string
+		configured string
+		expected   string
+	}{
+		{
+			name:       "configured proper value 10s",
+			configured: "10s",
+			expected:   "10s",
+		},
+		{
+			name:       "configured value less than 200ms",
+			configured: "0s",
+			expected:   "disabled",
+		},
+		{
+			name:       "configured proper value 1s",
+			configured: "1s",
+			expected:   "1s",
+		},
+		{
+			name:       "configured proper value 1500ms",
+			configured: "1500ms",
+			expected:   "1.5s",
+		},
+		{
+			name:       "configured to value greater than 0s but lower than 200ms",
+			configured: "150ms",
+			expected:   "disabled",
+		},
+		{
+			name:       "configured invalid value that cannot be parsed",
+			configured: "invalid",
+			expected:   "error",
+		},
+		{
+			name:       "configured negative value",
+			configured: "-1s",
+			expected:   "disabled",
+		},
+	}
+
+	input := make(chan watch.Event)
+	defer close(input)
+
+	for _, tc := range tcs {
+		t.Run(tc.name, func(t *testing.T) {
+			d, err := New(input, tc.configured)
+
+			switch tc.expected {
+			case "disabled":
+				if err != nil {
+					t.Fatalf("failed to create debouncer with debounce time %q", tc.configured)
+				}
+				if d != nil {
+					t.Fatalf("debouncer was created but should be disabled for value %q", tc.configured)
+				}
+			case "error":
+				if err == nil {
+					t.Fatalf("debouncer was created but should error for value %q", tc.configured)
+				}
+			default:
+				if d == nil {
+					t.Fatalf("debouncer was not created for value %q", tc.configured)
+				}
+
+				if d.debounceTime.String() != tc.expected {
+					t.Fatalf("invalid debounce time %q was configured instead of expected %q", d.debounceTime.String(), tc.expected)
+				}
+			}
+
+		})
+	}
+}
+
+func TestStartStop(t *testing.T) {
+	t.Run("Run and stop the debouncer without issues", func(t *testing.T) {
+		input := make(chan watch.Event)
+		defer close(input)
+
+		expected := "200ms"
+
+		d, err := New(input, expected)
+
+		if err != nil {
+			t.Fatalf("failed to create debouncer with debounce time %q", expected)
+		}
+
+		if d.debounceTime.String() != expected {
+			t.Fatalf("invalid debounce time %q was configured instead of expected %q", d.debounceTime.String(), expected)
+		}
+
+		ctx, cancel := context.WithCancel(context.Background())
+
+		wg := sync.WaitGroup{}
+
+		wg.Go(func() {
+			if err := d.Start(ctx); err != nil {
+				t.Fatalf("debouncer error: %s", err.Error())
+			}
+		})
+
+		cancel()
+
+		timedOut := waitTimeout(&wg, time.Second*3)
+
+		if timedOut {
+			t.Fatal("debouncer was not closed before timeout")
+		}
+	})
+}
+
+func TestDebouncing(t *testing.T) {
+	tcs := []string{"endpointslices", "endpoints", "services"}
+
+	for _, tc := range tcs {
+		t.Run(fmt.Sprintf("Get the newest event as the only one when using %s", tc), func(t *testing.T) {
+			expected := "500ms"
+
+			fw := watch.NewFake()
+			defer fw.Stop()
+
+			d, err := New(fw.ResultChan(), expected)
+
+			if err != nil {
+				t.Fatalf("failed to create debouncer with debounce time %q", expected)
+			}
+
+			if d.debounceTime.String() != expected {
+				t.Fatalf("invalid debounce time %q was configured instead of expected %q", d.debounceTime.String(), expected)
+			}
+
+			ctx, cancel := context.WithCancel(context.Background())
+
+			wg := sync.WaitGroup{}
+
+			wg.Go(func() {
+				if err := d.Start(ctx); err != nil {
+					t.Fatalf("debouncer error: %s", err.Error())
+				}
+			})
+
+			numOfUpdates := 100
+
+			switch tc {
+			case "endpointslices":
+				epslice := &discoveryv1.EndpointSlice{
+					ObjectMeta: metav1.ObjectMeta{
+						Name:      "test",
+						Namespace: "test",
+					},
+					Endpoints: make([]discoveryv1.Endpoint, 1),
+				}
+
+				addrEpslices := []string{}
+
+				for i := range numOfUpdates {
+					addrEpslices = append(addrEpslices, strconv.Itoa(i))
+					epslice.Endpoints[0].Addresses = addrEpslices
+					fw.Add(epslice)
+				}
+			case "endpoints":
+				ep := &v1.Endpoints{ //nolint:staticcheck
+					ObjectMeta: metav1.ObjectMeta{
+						Name:      "test",
+						Namespace: "test",
+					},
+					Subsets: make([]v1.EndpointSubset, 1), //nolint:staticcheck
+				}
+
+				addrEp := []v1.EndpointAddress{}
+
+				for i := range numOfUpdates {
+					addrEp = append(addrEp, v1.EndpointAddress{IP: strconv.Itoa(i)})
+					ep.Subsets[0].Addresses = addrEp
+					fw.Add(ep)
+				}
+			case "services":
+				svc := &v1.Service{
+					ObjectMeta: metav1.ObjectMeta{
+						Name:      "test",
+						Namespace: "test",
+					},
+				}
+
+				svcPorts := []v1.ServicePort{}
+
+				for i := range numOfUpdates {
+					svcPorts = append(svcPorts, v1.ServicePort{Port: int32(i)})
+					svc.Spec.Ports = svcPorts
+					fw.Add(svc)
+				}
+
+			default:
+				t.Fatal("unknown test", "type", tc)
+			}
+
+			out := <-d.output
+
+			switch tc {
+			case "endpointslices":
+				outEps, ok := out.Object.(*discoveryv1.EndpointSlice)
+				if !ok {
+					t.Fatal("got different type of object than EndpointSlice, failed to cast")
+				}
+
+				if len(outEps.Endpoints[0].Addresses) != numOfUpdates {
+					t.Fatalf("expected to aggregate %d events, but got %d", numOfUpdates, len(outEps.Endpoints[0].Addresses))
+				}
+			case "endpoints":
+				outEps, ok := out.Object.(*v1.Endpoints) //nolint:staticcheck
+				if !ok {
+					t.Fatal("got different type of object than EndpointSlice, failed to cast")
+				}
+
+				if len(outEps.Subsets[0].Addresses) != numOfUpdates {
+					t.Fatalf("expected to aggregate %d events, but got %d", numOfUpdates, len(outEps.Subsets[0].Addresses))
+				}
+			case "services":
+				outSvc, ok := out.Object.(*v1.Service) //nolint:staticcheck
+				if !ok {
+					t.Fatal("got different type of object than EndpointSlice, failed to cast")
+				}
+
+				if len(outSvc.Spec.Ports) != numOfUpdates {
+					t.Fatalf("expected to aggregate %d events, but got %d", numOfUpdates, len(outSvc.Spec.Ports))
+				}
+			default:
+				t.Fatal("unknown test", "type", tc)
+			}
+
+			cancel()
+
+			timedOut := waitTimeout(&wg, time.Second*3)
+
+			if timedOut {
+				t.Fatal("debouncer was not closed before timeout")
+			}
+		})
+	}
+}
+
+func TestTypeChange(t *testing.T) {
+	tcs := []string{"endpointslices", "endpoints", "services"}
+
+	for _, tc := range tcs {
+		t.Run(fmt.Sprintf("Get the newest event as the only one when using %s", tc), func(t *testing.T) {
+			expected := "500ms"
+
+			fw := watch.NewFake()
+			defer fw.Stop()
+
+			d, err := New(fw.ResultChan(), expected)
+
+			if err != nil {
+				t.Fatalf("failed to create debouncer with debounce time %q", expected)
+			}
+
+			if d.debounceTime.String() != expected {
+				t.Fatalf("invalid debounce time %q was configured instead of expected %q", d.debounceTime.String(), expected)
+			}
+
+			ctx, cancel := context.WithCancel(context.Background())
+
+			wg := sync.WaitGroup{}
+
+			wg.Go(func() {
+				if err := d.Start(ctx); err != nil {
+					t.Fatalf("debouncer error: %s", err.Error())
+				}
+			})
+
+			numOfUpdates := 100
+
+			switch tc {
+			case "endpointslices":
+				epslice := &discoveryv1.EndpointSlice{
+					ObjectMeta: metav1.ObjectMeta{
+						Name:      "test",
+						Namespace: "test",
+					},
+					Endpoints: make([]discoveryv1.Endpoint, 1),
+				}
+
+				addrEpslices := []string{}
+
+				for i := range numOfUpdates {
+					addrEpslices = append(addrEpslices, strconv.Itoa(i))
+					epslice.Endpoints[0].Addresses = addrEpslices
+					if i < numOfUpdates-1 {
+						fw.Add(epslice)
+					} else {
+						fw.Delete(epslice)
+					}
+				}
+			case "endpoints":
+				ep := &v1.Endpoints{ //nolint:staticcheck
+					ObjectMeta: metav1.ObjectMeta{
+						Name:      "test",
+						Namespace: "test",
+					},
+					Subsets: make([]v1.EndpointSubset, 1), //nolint:staticcheck
+				}
+
+				addrEp := []v1.EndpointAddress{}
+
+				for i := range numOfUpdates {
+					addrEp = append(addrEp, v1.EndpointAddress{IP: strconv.Itoa(i)})
+					ep.Subsets[0].Addresses = addrEp
+					if i < numOfUpdates-1 {
+						fw.Add(ep)
+					} else {
+						fw.Delete(ep)
+					}
+				}
+			case "services":
+				svc := &v1.Service{
+					ObjectMeta: metav1.ObjectMeta{
+						Name:      "test",
+						Namespace: "test",
+					},
+				}
+
+				svcPorts := []v1.ServicePort{}
+
+				for i := range numOfUpdates {
+					svcPorts = append(svcPorts, v1.ServicePort{Port: int32(i)})
+					svc.Spec.Ports = svcPorts
+					if i < numOfUpdates-1 {
+						fw.Add(svc)
+					} else {
+						fw.Delete(svc)
+					}
+				}
+
+			default:
+				t.Fatal("unknown test", "type", tc)
+			}
+
+			out := <-d.output
+
+			if out.Type != watch.Added {
+				t.Fatalf("expected to get add event, but got %s event", out.Type)
+			}
+
+			out = <-d.output
+
+			if out.Type != watch.Deleted {
+				t.Fatalf("expected to get delete event, but got %s event", out.Type)
+			}
+
+			cancel()
+
+			timedOut := waitTimeout(&wg, time.Second*3)
+
+			if timedOut {
+				t.Fatal("debouncer was not closed before timeout")
+			}
+		})
+	}
+}
+
+// waitTimeout waits for the waitgroup for the specified max timeout.
+// Returns true if waiting timed out.
+func waitTimeout(wg *sync.WaitGroup, timeout time.Duration) bool {
+	c := make(chan struct{})
+	go func() {
+		defer close(c)
+		wg.Wait()
+	}()
+	select {
+	case <-c:
+		return false // completed normally
+	case <-time.After(timeout):
+		return true // timed out
+	}
+}

pkg/detector/interfaces.go

@@ -27,7 +27,7 @@ func FindIPAddress(addrName string) (string, string, error) {
 					// If we're not searching for a specific adapter return the first one
 					if addrName == "" {
 						return iface.Name, address, nil
-					} else
+					}
 					// If this is the correct adapter return the details
 					if iface.Name == addrName {
 						return iface.Name, address, nil
@@ -37,5 +37,5 @@ func FindIPAddress(addrName string) (string, string, error) {
 		}
 
 	}
-	return "", "", fmt.Errorf("Unknown interface [%s]", addrName)
+	return "", "", fmt.Errorf("unknown interface [%s]", addrName)
 }

pkg/egress/egress.go

@@ -7,6 +7,7 @@ import (
 	"github.com/kube-vip/kube-vip/pkg/iptables"
 	"github.com/kube-vip/kube-vip/pkg/kubevip"
 	"github.com/kube-vip/kube-vip/pkg/nftables"
+	"github.com/kube-vip/kube-vip/pkg/utils"
 	"github.com/kube-vip/kube-vip/pkg/vip"
 )
 
@@ -18,15 +19,13 @@ func Teardown(podIP, vipIP, namespace, serviceUUID string, annotations map[strin
 	internalEgress := annotations[kubevip.EgressInternal]
 
 	protocol := iptables.ProtocolIPv4
-	IPv6 := false
-	if vip.IsIPv6(podIP) {
+	if utils.IsIPv6(podIP) {
 		protocol = iptables.ProtocolIPv6
-		IPv6 = true
 	}
 
 	// Use the internal egress implementation
-	if internalEgress != "" {
-		return nftables.DeleteSNAT(IPv6, serviceUUID)
+	if internalEgress != "" || useNftables {
+		return nftables.DeleteSNATFromAllTables(serviceUUID)
 	}
 
 	i, err := vip.CreateIptablesClient(useNftables, namespace, protocol)

pkg/election/election.go

@@ -0,0 +1,261 @@
+package election
+
+import (
+	"context"
+	"fmt"
+	"sync"
+	"time"
+
+	log "log/slog"
+
+	"github.com/kube-vip/kube-vip/pkg/etcd"
+	"github.com/kube-vip/kube-vip/pkg/kubevip"
+	"github.com/kube-vip/kube-vip/pkg/lease"
+	"github.com/kube-vip/kube-vip/pkg/loadbalancer"
+	"github.com/kube-vip/kube-vip/pkg/utils"
+	clientv3 "go.etcd.io/etcd/client/v3"
+	v1 "k8s.io/api/core/v1"
+	metav1 "k8s.io/apimachinery/pkg/apis/meta/v1"
+	"k8s.io/apimachinery/pkg/watch"
+	"k8s.io/client-go/kubernetes"
+	"k8s.io/client-go/tools/cache"
+	"k8s.io/client-go/tools/leaderelection"
+	"k8s.io/client-go/tools/leaderelection/resourcelock"
+	watchtools "k8s.io/client-go/tools/watch"
+)
+
+type Manager struct {
+	KubernetesClient   *kubernetes.Clientset
+	RetryWatcherClient *kubernetes.Clientset
+	// This channel is used to signal a shutdown
+
+	EtcdClient *clientv3.Client
+}
+
+// NewManager will create a new managing object
+func NewManager(config *kubevip.Config, k8sClientset, rwClientset *kubernetes.Clientset) (*Manager, error) {
+	m := &Manager{}
+
+	switch config.LeaderElectionType {
+	case "kubernetes", "":
+		if k8sClientset == nil || rwClientset == nil {
+			return nil, fmt.Errorf("provided nil clientset")
+		}
+		m.KubernetesClient = k8sClientset
+		m.RetryWatcherClient = rwClientset
+	case "etcd":
+		client, err := etcd.NewClient(config)
+		if err != nil {
+			return nil, err
+		}
+		m.EtcdClient = client
+	default:
+		return nil, fmt.Errorf("invalid LeaderElectionMode %s not supported", config.LeaderElectionType)
+	}
+
+	return m, nil
+}
+
+func RunOrDie(ctx context.Context, run *RunConfig, c *kubevip.Config) error {
+	switch c.LeaderElectionType {
+	case "kubernetes", "":
+		runKubernetesLeaderElectionOrDie(ctx, run)
+	case "etcd":
+		if err := runEtcdLeaderElectionOrDie(ctx, run); err != nil {
+			return err
+		}
+	default:
+		log.Info("LeaderElectionMode not supported, exiting", "mode", c.LeaderElectionType)
+	}
+
+	return nil
+}
+
+func runKubernetesLeaderElectionOrDie(ctx context.Context, run *RunConfig) {
+	// we use the Lease lock type since edits to Leases are less common
+	// and fewer objects in the cluster watch "all Leases".
+	lock := &resourcelock.LeaseLock{
+		LeaseMeta: metav1.ObjectMeta{
+			Name:        run.LeaseID.Name(),
+			Namespace:   run.LeaseID.Namespace(),
+			Annotations: run.LeaseAnnotations,
+		},
+		Client: run.Mgr.KubernetesClient.CoordinationV1(),
+		LockConfig: resourcelock.ResourceLockConfig{
+			Identity: run.Config.NodeName,
+		},
+	}
+
+	// start the leader election code loop
+	leaderelection.RunOrDie(ctx, leaderelection.LeaderElectionConfig{
+		Lock: lock,
+		// IMPORTANT: you MUST ensure that any code you have that
+		// is protected by the lease must terminate **before**
+		// you call cancel. Otherwise, you could have a background
+		// loop still running and another process could
+		// get elected before your background loop finished, violating
+		// the stated goal of the lease.
+		ReleaseOnCancel: true,
+		LeaseDuration:   time.Duration(run.Config.LeaseDuration) * time.Second,
+		RenewDeadline:   time.Duration(run.Config.RenewDeadline) * time.Second,
+		RetryPeriod:     time.Duration(run.Config.RetryPeriod) * time.Second,
+		Callbacks: leaderelection.LeaderCallbacks{
+			OnStartedLeading: run.OnStartedLeading,
+			OnStoppedLeading: run.OnStoppedLeading,
+			OnNewLeader:      run.OnNewLeader,
+		},
+	})
+}
+
+func runEtcdLeaderElectionOrDie(ctx context.Context, run *RunConfig) error {
+	if err := etcd.RunElectionOrDie(ctx, &etcd.LeaderElectionConfig{
+		EtcdConfig:           etcd.ClientConfig{Client: run.Mgr.EtcdClient},
+		Name:                 run.LeaseID.NamespacedName(),
+		MemberID:             run.Config.NodeName,
+		LeaseDurationSeconds: int64(run.Config.LeaseDuration),
+		Callbacks: etcd.LeaderCallbacks{
+			OnStartedLeading: run.OnStartedLeading,
+			OnStoppedLeading: run.OnStoppedLeading,
+			OnNewLeader:      run.OnNewLeader,
+		},
+	}); err != nil {
+		return fmt.Errorf("etcd leaderelection: %w", err)
+	}
+	return nil
+}
+
+type Actions interface {
+	OnStartedLeading(ctx context.Context)
+	OnStoppedLeading()
+	OnNewLeader(identity string)
+}
+
+type RunConfig struct {
+	Config           *kubevip.Config
+	LeaseID          lease.ID
+	Mgr              *Manager
+	LeaseAnnotations map[string]string
+
+	// onStartedLeading is called when this member starts leading.
+	OnStartedLeading func(context.Context)
+	// onStoppedLeading is called when this member stops leading.
+	OnStoppedLeading func()
+	// onNewLeader is called when the client observes a leader that is
+	// not the previously observed leader. This includes the first observed
+	// leader when the client starts.
+	OnNewLeader func(identity string)
+}
+
+func (em *Manager) NodeWatcher(ctx context.Context, lb *loadbalancer.IPVSLoadBalancer, port uint16) error {
+	// Use a restartable watcher, as this should help in the event of etcd or timeout issues
+	log.Info("Kube-Vip is watching nodes for control-plane labels")
+
+	listOptions := metav1.ListOptions{
+		LabelSelector: "node-role.kubernetes.io/control-plane",
+	}
+
+	wg := sync.WaitGroup{}
+	defer wg.Wait()
+
+	watchCtx, watchCancel := context.WithCancel(ctx)
+	defer watchCancel()
+
+	rw, err := watchtools.NewRetryWatcherWithContext(watchCtx, "1", &cache.ListWatch{
+		WatchFunc: func(_ metav1.ListOptions) (watch.Interface, error) {
+			return utils.WatchWithAuthRetry(ctx, func(ctx context.Context) (watch.Interface, error) {
+				return em.RetryWatcherClient.CoreV1().Nodes().Watch(watchCtx, listOptions)
+			})
+		},
+	})
+	if err != nil {
+		return fmt.Errorf("error creating label watcher: %w", err)
+	}
+
+	wg.Go(func() {
+		<-watchCtx.Done()
+		log.Info("Node watcher context cancelled, stopping")
+		// Stop the retrywatcher
+		rw.Stop()
+	})
+
+	ch := rw.ResultChan()
+
+	var watchErr error
+	for event := range ch {
+		// We need to inspect the event and get ResourceVersion out of it
+		switch event.Type {
+		case watch.Added, watch.Modified:
+			node, ok := event.Object.(*v1.Node)
+			if !ok {
+				return fmt.Errorf("unable to parse Kubernetes Node from Annotation watcher")
+			}
+			// Find the node IP address (this isn't foolproof)
+			for x := range node.Status.Addresses {
+				if node.Status.Addresses[x].Type == v1.NodeInternalIP {
+					if checkIfNodeIsReady(node) {
+						err = lb.AddBackend(node.Status.Addresses[x].Address, port)
+						if err != nil {
+							log.Error("adding node to load balancer", "node", node.Name, "ip", node.Status.Addresses[x].Address, "err", err)
+							if utils.IsPanicError(err) {
+								return fmt.Errorf("add IPVS backend: %w", err)
+							}
+						}
+					} else {
+						err = lb.RemoveBackend(node.Status.Addresses[x].Address, port)
+						if err != nil {
+							log.Error("removing node from load balancer", "node", node.Name, "ip", node.Status.Addresses[x].Address, "err", err)
+						}
+					}
+				}
+			}
+		case watch.Deleted:
+			node, ok := event.Object.(*v1.Node)
+			if !ok {
+				return fmt.Errorf("unable to parse Kubernetes Node from Annotation watcher")
+			}
+
+			// Find the node IP address (this isn't foolproof)
+			for x := range node.Status.Addresses {
+				if node.Status.Addresses[x].Type == v1.NodeInternalIP {
+					err = lb.RemoveBackend(node.Status.Addresses[x].Address, port)
+					if err != nil {
+						log.Error("removing node from load balancer", "node", node.Name, "ip", node.Status.Addresses[x].Address, "err", err)
+					}
+				}
+			}
+
+			log.Info("Node deleted", "name", node.Name)
+
+		case watch.Bookmark:
+			// Un-used
+		case watch.Error:
+			log.Error("Error attempting to watch Kubernetes Nodes")
+			watchErr = fmt.Errorf("node watcher error: %w", utils.WatchError(event.Object))
+			log.Error("watcher", "err", watchErr)
+		default:
+		}
+	}
+
+	log.Info("Exiting Node watcher")
+	if watchErr != nil {
+		return watchErr
+	}
+	if ctx.Err() != nil {
+		return nil
+	}
+	return utils.NewPanicError("node watcher channel closed unexpectedly")
+}
+
+func checkIfNodeIsReady(node *v1.Node) bool {
+	if node == nil {
+		return false
+	}
+	for _, condition := range node.Status.Conditions {
+		if condition.Type == v1.NodeReady {
+			if condition.Status == v1.ConditionTrue {
+				return true
+			}
+		}
+	}
+	return false
+}

pkg/endpoints/endpoints.go

@@ -4,6 +4,9 @@ import (
 	"context"
 	"fmt"
 	"net"
+	"strings"
+	"sync"
+	"time"
 
 	log "log/slog"
 
@@ -11,10 +14,15 @@ import (
 	"github.com/kube-vip/kube-vip/pkg/endpoints/providers"
 	"github.com/kube-vip/kube-vip/pkg/instance"
 	"github.com/kube-vip/kube-vip/pkg/kubevip"
+	"github.com/kube-vip/kube-vip/pkg/lease"
+	"github.com/kube-vip/kube-vip/pkg/metrics"
+	"github.com/kube-vip/kube-vip/pkg/route"
 	"github.com/kube-vip/kube-vip/pkg/servicecontext"
+	"github.com/kube-vip/kube-vip/pkg/utils"
+	"github.com/kube-vip/kube-vip/pkg/wireguard"
 	v1 "k8s.io/api/core/v1"
-	discoveryv1 "k8s.io/api/discovery/v1"
 	"k8s.io/apimachinery/pkg/watch"
+	"k8s.io/client-go/kubernetes"
 )
 
 type Processor struct {
@@ -23,26 +31,29 @@ type Processor struct {
 	bgpServer *bgp.Server
 	worker    endpointWorker
 	instances *[]*instance.Instance
+	leaseMgr  *lease.Manager
 }
 
 func NewEndpointProcessor(config *kubevip.Config, provider providers.Provider, bgpServer *bgp.Server,
-	instances *[]*instance.Instance) *Processor {
+	instances *[]*instance.Instance, leaseMgr *lease.Manager, tunnelMgr *wireguard.TunnelManager, routeMgr *route.Manager) *Processor {
 	return &Processor{
 		config:    config,
 		provider:  provider,
 		bgpServer: bgpServer,
 		instances: instances,
-		worker:    newEndpointWorker(config, provider, bgpServer, instances),
+		leaseMgr:  leaseMgr,
+		worker:    newEndpointWorker(config, provider, bgpServer, instances, leaseMgr, tunnelMgr, routeMgr),
 	}
 }
 
-func (p *Processor) AddOrModify(ctx *servicecontext.Context, event watch.Event,
-	lastKnownGoodEndpoint *string, service *v1.Service, id string, leaderElectionActive *bool,
-	serviceFunc func(context.Context, *v1.Service) error,
-	leaderCtx *context.Context, cancel *context.CancelFunc) (bool, error) {
+func (p *Processor) AddOrModify(svcCtx *servicecontext.Context, event watch.Event,
+	lastKnownGoodEndpoint *string, service *v1.Service, id string,
+	serviceFunc func(*servicecontext.Context, *v1.Service, *sync.WaitGroup, bool) error, wg *sync.WaitGroup,
+	clientSet *kubernetes.Clientset,
+	egressUpdateFunc func(context.Context, *v1.Service) error) (bool, error) {
 
 	var err error
-	if err = p.provider.LoadObject(event.Object, *cancel); err != nil {
+	if err = p.provider.LoadObject(event.Object, svcCtx.Cancel); err != nil {
 		return false, fmt.Errorf("[%s] error loading k8s object: %w", p.provider.GetLabel(), err)
 	}
 
@@ -51,10 +62,12 @@ func (p *Processor) AddOrModify(ctx *servicecontext.Context, event watch.Event,
 		return false, err
 	}
 
-	if err := p.worker.setInstanceEndpointsStatus(service, endpoints); err != nil {
+	if err := p.worker.setInstanceEndpointsStatus(svcCtx.Ctx, service, endpoints); err != nil {
 		log.Error("updating instance", "err", err)
 	}
 
+	allowReconcileWithoutEndpoints := shouldAllowReconcileWithoutEndpoints(service)
+
 	// Find out if we have any local endpoints
 	// if out endpoint is empty then populate it
 	// if not, go through the endpoints and see if ours still exists
@@ -64,50 +77,80 @@ func (p *Processor) AddOrModify(ctx *servicecontext.Context, event watch.Event,
 	// Check that we have local endpoints
 	if len(endpoints) != 0 {
 		// Ignore IPv4
-		if service.Annotations[kubevip.EgressIPv6] == "true" && net.ParseIP(endpoints[0]).To4() != nil {
+		if service.Annotations[kubevip.EgressIPv6] == "true" && !hasV6(endpoints) {
 			return true, nil
 		}
 
-		p.updateLastKnownGoodEndpoint(lastKnownGoodEndpoint, endpoints, service, leaderElectionActive, *cancel)
-		// start leader election if it's enabled and not already started
-		if !*leaderElectionActive && p.config.EnableServicesElection {
-			go func() {
-				*leaderCtx, *cancel = context.WithCancel(ctx.Ctx)
-				startLeaderElection(*leaderCtx, leaderElectionActive, service, serviceFunc)
-			}()
+		p.updateLastKnownGoodEndpoint(lastKnownGoodEndpoint, endpoints, service)
+
+		if err := p.startServiceHandlingIfNeeded(svcCtx, service, serviceFunc, wg); err != nil {
+			return true, err
 		}
 
+		svcCtx.SignalReadiness()
+
 		// There are local endpoints available on the node
-		if !p.config.EnableServicesElection && !p.config.EnableLeaderElection {
-			if err := p.worker.processInstance(ctx, service, leaderElectionActive); err != nil {
+		// Process immediately if:
+		// - No services/leader election is enabled, OR
+		// - WireGuard is enabled (it always needs immediate DNAT rule updates)
+		if (!p.config.EnableServicesElection && !p.config.EnableLeaderElection) || p.config.EnableWireguard {
+			if err := p.worker.processInstance(svcCtx, service); err != nil {
 				return false, fmt.Errorf("failed to process non-empty instance: %w", err)
 			}
 		}
 	} else {
-		// There are no local endpoints
-		p.worker.clear(ctx, lastKnownGoodEndpoint, service, *cancel, leaderElectionActive)
+		if allowReconcileWithoutEndpoints {
+			// Explicit opt-in for controllers that create LoadBalancer services without endpoints
+			if err := p.startServiceHandlingIfNeeded(svcCtx, service, serviceFunc, wg); err != nil {
+				return true, err
+			}
+			svcCtx.SignalReadiness()
+
+			if (!p.config.EnableServicesElection && !p.config.EnableLeaderElection) || p.config.EnableWireguard {
+				if err := p.worker.processInstance(svcCtx, service); err != nil {
+					return false, fmt.Errorf("failed to process endpointless instance: %w", err)
+				}
+			}
+		} else if svcCtx.Signalled.Load() {
+			// There are no local endpoints
+			svcCtx.ResetReadiness()
+			p.worker.clear(svcCtx, lastKnownGoodEndpoint, service)
+			if p.config.EnableARP && !p.config.EnableServicesElection {
+				i := instance.FindServiceInstance(service, *p.instances)
+				for _, c := range i.Clusters {
+					c.Stop()
+				}
+			}
+		}
 	}
 
 	// Set the service accordingly
-	p.updateAnnotations(service, lastKnownGoodEndpoint)
+	p.updateAnnotations(service, lastKnownGoodEndpoint, clientSet, egressUpdateFunc)
 
 	log.Debug("watcher", "provider",
-		p.provider.GetLabel(), "service name", service.Name, "namespace", service.Namespace, "endpoints", len(endpoints), "last endpoint", *lastKnownGoodEndpoint, "active leader election", *leaderElectionActive)
+		p.provider.GetLabel(), "service name", service.Name, "namespace", service.Namespace, "endpoints", len(endpoints), "last endpoint", *lastKnownGoodEndpoint)
 
 	return false, nil
 }
 
-func (p *Processor) Delete(service *v1.Service, id string) error {
-	if err := p.worker.delete(service, id); err != nil {
+func (p *Processor) Delete(ctx context.Context, service *v1.Service, id string) error {
+	if err := p.worker.delete(ctx, service, id); err != nil {
 		return fmt.Errorf("[%s] error deleting service: %w", p.provider.GetLabel(), err)
 	}
 	return nil
 }
 
-func (p *Processor) updateLastKnownGoodEndpoint(lastKnownGoodEndpoint *string, endpoints []string, service *v1.Service, leaderElectionActive *bool, cancel context.CancelFunc) {
+func (p *Processor) updateLastKnownGoodEndpoint(lastKnownGoodEndpoint *string, endpoints []string, service *v1.Service) {
 	// if we haven't populated one, then do so
+	family := utils.IPv4Family
+	if service.Annotations[kubevip.EgressIPv6] == "true" {
+		family = utils.IPv6Family
+	}
+
+	ep := getEndpoint(endpoints, family)
+
 	if *lastKnownGoodEndpoint == "" {
-		*lastKnownGoodEndpoint = endpoints[0]
+		*lastKnownGoodEndpoint = ep
 		return
 	}
 
@@ -121,45 +164,197 @@ func (p *Processor) updateLastKnownGoodEndpoint(lastKnownGoodEndpoint *string, e
 	}
 	// If the last endpoint no longer exists, we cancel our leader Election, and set another endpoint as last known good
 	if !stillExists {
-		p.worker.removeEgress(service, lastKnownGoodEndpoint)
-		if *leaderElectionActive && (p.config.EnableServicesElection || p.config.EnableLeaderElection) {
-			log.Warn("existing endpoint has been removed, restarting leaderElection", "provider", p.provider.GetLabel(), "endpoint", *lastKnownGoodEndpoint)
-			// Stop the existing leaderElection
-			cancel()
-			// disable last leaderElection flag
-			*leaderElectionActive = false
+		ip := net.ParseIP(*lastKnownGoodEndpoint)
+		if (ip.To4() != nil && service.Annotations[kubevip.Egress] == "true") ||
+			(ip.To4() == nil && service.Annotations[kubevip.EgressIPv6] == "true") {
+			p.worker.removeEgress(service, lastKnownGoodEndpoint)
 		}
 		// Set our active endpoint to an existing one
-		*lastKnownGoodEndpoint = endpoints[0]
+		*lastKnownGoodEndpoint = ep
 	}
 }
 
-func (p *Processor) updateAnnotations(service *v1.Service, lastKnownGoodEndpoint *string) {
+func (p *Processor) updateAnnotations(service *v1.Service, lastKnownGoodEndpoint *string,
+	clientSet *kubernetes.Clientset,
+	egressUpdateFunc func(context.Context, *v1.Service) error) {
 	// Set the service accordingly
 	if service.Annotations[kubevip.Egress] == "true" {
-		activeEndpointAnnotation := kubevip.ActiveEndpoint
-
-		if !p.config.EnableEndpoints && p.provider.GetProtocol() == string(discoveryv1.AddressTypeIPv6) {
-			activeEndpointAnnotation = kubevip.ActiveEndpointIPv6
+		if *lastKnownGoodEndpoint != "" {
+			ip := net.ParseIP(*lastKnownGoodEndpoint)
+			expectIPv6 := service.Annotations[kubevip.EgressIPv6] == "true"
+			if ip == nil || (ip.To4() == nil) != expectIPv6 {
+				log.Warn("ignoring active endpoint with unexpected address family",
+					"service", service.Name,
+					"namespace", service.Namespace,
+					"endpoint", *lastKnownGoodEndpoint,
+					"expected_ipv6", expectIPv6)
+				return
+			}
+		}
+
+		// Store old values from ServiceSnapshot to detect if annotation actually changed
+		// We use the ServiceSnapshot instead of the service parameter because the service parameter
+		// may have stale annotations if the last update failed
+		var oldEndpoint, oldEndpointIPv6 string
+		snapshotFound := false
+		if p.instances != nil {
+			serviceInstance := instance.FindServiceInstance(service, *p.instances)
+			if serviceInstance != nil && serviceInstance.ServiceSnapshot != nil {
+				snapshotFound = true
+				oldEndpoint = serviceInstance.ServiceSnapshot.Annotations[kubevip.ActiveEndpoint]
+				oldEndpointIPv6 = serviceInstance.ServiceSnapshot.Annotations[kubevip.ActiveEndpointIPv6]
+			}
+		}
+		// Empty annotations in an existing snapshot are meaningful after a zero-endpoint transition.
+		if !snapshotFound {
+			oldEndpoint = service.Annotations[kubevip.ActiveEndpoint]
+			oldEndpointIPv6 = service.Annotations[kubevip.ActiveEndpointIPv6]
+		}
+
+		// Determine which annotation to update based on IP version
+		var endpoint, endpointIPv6 string
+		if service.Annotations[kubevip.EgressIPv6] == "true" && !p.config.EnableEndpoints {
+			// IPv6
+			endpointIPv6 = *lastKnownGoodEndpoint
+			endpoint = oldEndpoint // Preserve existing IPv4 if any
+		} else {
+			// IPv4
+			endpoint = *lastKnownGoodEndpoint
+			endpointIPv6 = oldEndpointIPv6 // Preserve existing IPv6 if any
+		}
+
+		// Check if annotation actually changed
+		annotationChanged := (oldEndpoint != endpoint) || (oldEndpointIPv6 != endpointIPv6)
+		if !annotationChanged {
+			return // Nothing to do
+		}
+
+		// Persist to Kubernetes
+		ctx := context.Background()
+
+		if err := p.provider.UpdateServiceAnnotation(ctx, endpoint, endpointIPv6, service, clientSet); err != nil {
+			log.Warn("failed to update service annotation", "service", service.Name, "namespace", service.Namespace, "err", err)
+			return
+		}
+
+		log.Debug("updated active endpoint annotation", "service", service.Name, "namespace", service.Namespace, "endpoint", *lastKnownGoodEndpoint)
+
+		// Trigger egress reconfiguration
+		// For services with leader election, the service watcher doesn't process Modified events
+		// after initial setup, so we need to directly call the update function
+		if egressUpdateFunc != nil {
+			// Create a copy of service with updated annotations
+			svcCopy := service.DeepCopy()
+			svcCopy.Annotations[kubevip.ActiveEndpoint] = endpoint
+			svcCopy.Annotations[kubevip.ActiveEndpointIPv6] = endpointIPv6
+
+			if err := egressUpdateFunc(ctx, svcCopy); err != nil {
+				log.Error("failed to reconfigure egress", "service", service.Name, "namespace", service.Namespace, "err", err)
+			}
 		}
-		service.Annotations[activeEndpointAnnotation] = *lastKnownGoodEndpoint
 	}
 }
 
-func startLeaderElection(ctx context.Context, leaderElectionActive *bool, service *v1.Service, serviceFunc func(context.Context, *v1.Service) error) {
+func (p *Processor) startServiceHandlingIfNeeded(svcCtx *servicecontext.Context, service *v1.Service,
+	serviceFunc func(*servicecontext.Context, *v1.Service, *sync.WaitGroup, bool) error, wg *sync.WaitGroup) error {
+	if p.config.EnableServicesElection {
+		// startLeaderElection restarts itself until the service context is cancelled,
+		// so start it only once instead of on every endpoint event.
+		svcCtx.StartLeaderElectionOnce(func() {
+			wg.Go(func() {
+				p.startLeaderElection(svcCtx, service, serviceFunc, wg)
+			})
+		})
+		return nil
+	}
+
+	if p.config.EnableARP || (p.config.EnableRoutingTable && p.config.EnableLeaderElection) {
+		if !svcCtx.Signalled.Load() {
+			inst := instance.FindServiceInstance(service, *p.instances)
+			if inst == nil {
+				return fmt.Errorf("[%s] failed to find an instance for service %s/%s", p.provider.GetLabel(), service.Namespace, service.Name)
+			}
+			for x := range inst.VIPConfigs {
+				log.Debug("starting loadbalancer for service", "provider", p.provider.GetLabel(), "name", service.Name, "namespace", service.Namespace, "uid", service.UID)
+				if err := inst.Clusters[x].StartLoadBalancerService(svcCtx.Ctx, inst.VIPConfigs[x], p.bgpServer, lease.ServiceNamespacedName(service), wg); err != nil {
+					return fmt.Errorf("failed to start lb: %w", err)
+				}
+			}
+		}
+	}
+
+	return nil
+}
+
+func (p *Processor) startLeaderElection(svcCtx *servicecontext.Context, service *v1.Service, serviceFunc func(*servicecontext.Context, *v1.Service, *sync.WaitGroup, bool) error, wg *sync.WaitGroup) {
+	// Track this loop for the lifetime of the goroutine. There has to be at most
+	// one per service, so a value above 1 means loops leaked.
+	loops := metrics.ServiceElectionLoops.WithLabelValues(service.Namespace, service.Name)
+	loops.Inc()
+	defer loops.Dec()
+
+	attempts := metrics.ServiceElectionAttemptsTotal.WithLabelValues(service.Namespace, service.Name)
+
 	// This is a blocking function, that will restart (in the event of failure)
 	for {
-		// if the context isn't cancelled restart
-		if ctx.Err() != context.Canceled {
-			*leaderElectionActive = true
-			err := serviceFunc(ctx, service)
-			if err != nil {
-				log.Error(err.Error())
+		select {
+		case <-svcCtx.Ctx.Done():
+			return
+		default:
+			leaseNamespace, serviceLease := lease.ServiceName(service)
+			id := lease.NewID(p.config.LeaderElectionType, leaseNamespace, serviceLease)
+			// The lease is retired once its last service is gone, so an absent one means
+			// this loop has nothing left to elect for.
+			l := p.leaseMgr.Get(id)
+			if l == nil {
+				return
+			}
+			l.Lock()
+
+			if !l.Elected.Load() {
+				l.Unlock()
+				attempts.Inc()
+				err := serviceFunc(svcCtx, service, wg, true)
+				if err != nil {
+					log.Error(err.Error())
+				}
+			} else {
+				l.Unlock()
+				time.Sleep(time.Millisecond * 200)
 			}
-			*leaderElectionActive = false
-		} else {
-			*leaderElectionActive = false
-			break
 		}
 	}
 }
+
+func shouldAllowReconcileWithoutEndpoints(service *v1.Service) bool {
+	if service == nil || service.Spec.ExternalTrafficPolicy != v1.ServiceExternalTrafficPolicyTypeCluster {
+		return false
+	}
+
+	return strings.EqualFold(service.Annotations[kubevip.AllowReconcileWithoutEndpoints], "true")
+}
+
+func hasV6(endpoints []string) bool {
+	for _, e := range endpoints {
+		ip := net.ParseIP(e)
+		if ip != nil {
+			if ip.To4() == nil {
+				return true
+			}
+		}
+	}
+	return false
+}
+
+func getEndpoint(endpoints []string, family string) string {
+	for _, e := range endpoints {
+		ip := net.ParseIP(e)
+		if family == utils.IPv4Family && ip.To4() != nil {
+			return e
+		}
+		if family == utils.IPv6Family && ip.To4() == nil {
+			return e
+		}
+	}
+	return ""
+}

pkg/endpoints/endpoints_bgp.go

@@ -7,6 +7,7 @@ import (
 
 	"github.com/kube-vip/kube-vip/pkg/bgp"
 	"github.com/kube-vip/kube-vip/pkg/instance"
+	"github.com/kube-vip/kube-vip/pkg/lease"
 	"github.com/kube-vip/kube-vip/pkg/servicecontext"
 	v1 "k8s.io/api/core/v1"
 )
@@ -23,20 +24,19 @@ func newBGP(generic generic, bgpServer *bgp.Server) endpointWorker {
 	}
 }
 
-func (b *BGP) processInstance(ctx *servicecontext.Context, service *v1.Service, leaderElectionActive *bool) error {
+func (b *BGP) processInstance(svcCtx *servicecontext.Context, service *v1.Service) error {
 	if instance := instance.FindServiceInstance(service, *b.instances); instance != nil {
 		for _, cluster := range instance.Clusters {
 			for i := range cluster.Network {
-				if !ctx.IsNetworkConfigured(cluster.Network[i].IP()) {
+				if !svcCtx.IsNetworkConfigured(cluster.Network[i].IP()) {
 					log.Debug("attempting to advertise BGP service", "provider", b.provider.GetLabel(), "ip", cluster.Network[i].IP())
-					err := b.bgpServer.AddHost(cluster.Network[i].CIDR())
+					err := b.bgpServer.AddHost(svcCtx.Ctx, cluster.Network[i].CIDR(), lease.ServiceNamespacedName(service))
 					if err != nil {
 						log.Error("error adding BGP host", "provider", b.provider.GetLabel(), "err", err)
 					} else {
 						log.Info("added BGP host", "provider",
 							b.provider.GetLabel(), "ip", cluster.Network[i].CIDR(), "service name", service.Name, "namespace", service.Namespace)
-						ctx.ConfiguredNetworks.Store(cluster.Network[i].IP(), true)
-						*leaderElectionActive = true
+						svcCtx.ConfiguredNetworks.Store(cluster.Network[i].IP(), true)
 					}
 				}
 			}
@@ -45,20 +45,19 @@ func (b *BGP) processInstance(ctx *servicecontext.Context, service *v1.Service,
 	return nil
 }
 
-func (b *BGP) clear(ctx *servicecontext.Context, lastKnownGoodEndpoint *string, service *v1.Service, cancel context.CancelFunc, leaderElectionActive *bool) {
+func (b *BGP) clear(svcCtx *servicecontext.Context, lastKnownGoodEndpoint *string, service *v1.Service) {
 	if !b.config.EnableServicesElection && !b.config.EnableLeaderElection {
 		// If BGP mode is enabled - routes should be deleted
 		if instance := instance.FindServiceInstance(service, *b.instances); instance != nil {
 			for _, cluster := range instance.Clusters {
 				for i := range cluster.Network {
-					err := b.bgpServer.DelHost(cluster.Network[i].CIDR())
+					err := b.bgpServer.DelHost(svcCtx.Ctx, cluster.Network[i].CIDR(), lease.ServiceNamespacedName(service))
 					if err != nil {
 						log.Error("deleting BGP host", "provider", b.provider.GetLabel(), "ip", cluster.Network[i].IP(), "err", err)
 					} else {
 						log.Info("deleted BGP host", "provider",
 							b.provider.GetLabel(), "ip", cluster.Network[i].IP(), "service name", service.Name, "namespace", service.Namespace)
-						ctx.ConfiguredNetworks.Delete(cluster.Network[i])
-						*leaderElectionActive = false
+						svcCtx.ConfiguredNetworks.Delete(cluster.Network[i].IP())
 					}
 				}
 			}
@@ -66,14 +65,18 @@ func (b *BGP) clear(ctx *servicecontext.Context, lastKnownGoodEndpoint *string,
 		}
 	}
 
-	b.clearEgress(lastKnownGoodEndpoint, service, cancel, leaderElectionActive)
+	b.clearEgress(lastKnownGoodEndpoint, service)
+
+	if svcCtx.LeaderCancel != nil {
+		svcCtx.LeaderCancel()
+	}
 }
 
 func (b *BGP) getEndpoints(service *v1.Service, id string) ([]string, error) {
 	return b.getAllEndpoints(service, id)
 }
 
-func (b *BGP) delete(service *v1.Service, id string) error {
+func (b *BGP) delete(ctx context.Context, service *v1.Service, id string) error {
 	// When no-leader-elecition mode
 	if !b.config.EnableServicesElection && !b.config.EnableLeaderElection {
 		// find all existing local endpoints
@@ -84,36 +87,40 @@ func (b *BGP) delete(service *v1.Service, id string) error {
 
 		// If there were local endpoints deleted
 		if len(endpoints) > 0 {
-			b.deleteAction(service)
+			b.deleteAction(ctx, service)
 		}
 	}
 
 	return nil
 }
 
-func (b *BGP) deleteAction(service *v1.Service) {
-	b.clearBGPHosts(service)
+func (b *BGP) deleteAction(ctx context.Context, service *v1.Service) {
+	b.clearBGPHosts(ctx, service)
 }
 
-func (b *BGP) clearBGPHosts(service *v1.Service) {
-	ClearBGPHosts(service, b.instances, b.bgpServer)
+func (b *BGP) clearBGPHosts(ctx context.Context, service *v1.Service) {
+	ClearBGPHosts(ctx, service, b.instances, b.bgpServer)
 }
 
-func (b *BGP) setInstanceEndpointsStatus(_ *v1.Service, _ []string) error {
+func (b *BGP) setInstanceEndpointsStatus(_ context.Context, _ *v1.Service, _ []string) error {
 	return nil
 }
 
-func ClearBGPHosts(service *v1.Service, instances *[]*instance.Instance, bgpServer *bgp.Server) {
+func ClearBGPHosts(ctx context.Context, service *v1.Service, instances *[]*instance.Instance, bgpServer *bgp.Server) {
 	if instance := instance.FindServiceInstance(service, *instances); instance != nil {
-		ClearBGPHostsByInstance(instance, bgpServer)
+		ClearBGPHostsByInstance(ctx, instance, bgpServer)
 	}
 }
 
-func ClearBGPHostsByInstance(instance *instance.Instance, bgpServer *bgp.Server) {
+func ClearBGPHostsByInstance(ctx context.Context, instance *instance.Instance, bgpServer *bgp.Server) {
+	if instance == nil {
+		log.Error("failed to clear BGP host for nil instance")
+		return
+	}
 	for _, cluster := range instance.Clusters {
 		for i := range cluster.Network {
 			network := cluster.Network[i]
-			err := bgpServer.DelHost(network.CIDR())
+			err := bgpServer.DelHost(ctx, network.CIDR(), lease.ServiceNamespacedName(instance.ServiceSnapshot))
 			if err != nil {
 				log.Error("[endpoint] error deleting BGP host", "err", err)
 			} else {

pkg/endpoints/endpoints_generic.go

@@ -10,24 +10,31 @@ import (
 	"github.com/kube-vip/kube-vip/pkg/endpoints/providers"
 	"github.com/kube-vip/kube-vip/pkg/instance"
 	"github.com/kube-vip/kube-vip/pkg/kubevip"
+	"github.com/kube-vip/kube-vip/pkg/lease"
+	"github.com/kube-vip/kube-vip/pkg/route"
 	"github.com/kube-vip/kube-vip/pkg/servicecontext"
+	"github.com/kube-vip/kube-vip/pkg/wireguard"
 	v1 "k8s.io/api/core/v1"
 )
 
 type endpointWorker interface {
-	processInstance(svcCtx *servicecontext.Context, service *v1.Service, leaderElectionActive *bool) error
-	clear(svcCtx *servicecontext.Context, lastKnownGoodEndpoint *string, service *v1.Service, cancel context.CancelFunc, leaderElectionActive *bool)
+	processInstance(svcCtx *servicecontext.Context, service *v1.Service) error
+	clear(svcCtx *servicecontext.Context, lastKnownGoodEndpoint *string, service *v1.Service)
 	getEndpoints(service *v1.Service, id string) ([]string, error)
 	removeEgress(service *v1.Service, lastKnownGoodEndpoint *string)
-	delete(service *v1.Service, id string) error
-	setInstanceEndpointsStatus(service *v1.Service, endpoints []string) error
+	delete(ctx context.Context, service *v1.Service, id string) error
+	setInstanceEndpointsStatus(ctx context.Context, service *v1.Service, endpoints []string) error
 }
 
-func newEndpointWorker(config *kubevip.Config, provider providers.Provider, bgpServer *bgp.Server, instances *[]*instance.Instance) endpointWorker {
-	generic := newGeneric(config, provider, instances)
+func newEndpointWorker(config *kubevip.Config, provider providers.Provider, bgpServer *bgp.Server, instances *[]*instance.Instance,
+	leaseMgr *lease.Manager, tunnelMgr *wireguard.TunnelManager, routeMgr *route.Manager) endpointWorker {
+	generic := newGeneric(config, provider, instances, leaseMgr)
 
+	if config.EnableWireguard {
+		return newWireguardWorker(config, provider, bgpServer, instances, leaseMgr, tunnelMgr)
+	}
 	if config.EnableRoutingTable {
-		return newRoutingTable(generic)
+		return newRoutingTable(generic, routeMgr)
 	}
 	if config.EnableBGP {
 		return newBGP(generic, bgpServer)
@@ -40,60 +47,49 @@ type generic struct {
 	config    *kubevip.Config
 	provider  providers.Provider
 	instances *[]*instance.Instance
+	leaseMgr  *lease.Manager
 }
 
-func newGeneric(config *kubevip.Config, provider providers.Provider, instances *[]*instance.Instance) generic {
+func newGeneric(config *kubevip.Config, provider providers.Provider, instances *[]*instance.Instance, leaseMgr *lease.Manager) generic {
 	return generic{
 		config:    config,
 		provider:  provider,
 		instances: instances,
+		leaseMgr:  leaseMgr,
 	}
 }
 
-func (g *generic) processInstance(_ *servicecontext.Context, _ *v1.Service, _ *bool) error {
+func (g *generic) processInstance(_ *servicecontext.Context, _ *v1.Service) error {
 	return nil
 }
 
-func (g *generic) clear(_ *servicecontext.Context, lastKnownGoodEndpoint *string, service *v1.Service, cancel context.CancelFunc, leaderElectionActive *bool) {
-	g.clearEgress(lastKnownGoodEndpoint, service, cancel, leaderElectionActive)
+func (g *generic) clear(svcCtx *servicecontext.Context, lastKnownGoodEndpoint *string, service *v1.Service) {
+	g.clearEgress(lastKnownGoodEndpoint, service)
+	if svcCtx.LeaderCancel != nil {
+		svcCtx.LeaderCancel()
+	}
 }
 
-func (g *generic) clearEgress(lastKnownGoodEndpoint *string, service *v1.Service, cancel context.CancelFunc, leaderElectionActive *bool) {
+func (g *generic) clearEgress(lastKnownGoodEndpoint *string, service *v1.Service) {
 	if *lastKnownGoodEndpoint != "" {
-		log.Warn("existing  endpoint has been removed, no remaining endpoints for leaderElection", "provider", g.provider.GetLabel(), "endpoint", lastKnownGoodEndpoint)
+		log.Warn("existing endpoint has been removed, no remaining endpoints for leaderElection", "provider", g.provider.GetLabel(), "endpoint", lastKnownGoodEndpoint)
 		if err := egress.Teardown(*lastKnownGoodEndpoint, service.Spec.LoadBalancerIP, service.Namespace, string(service.UID), service.Annotations, g.config.EgressWithNftables); err != nil {
 			log.Error("error removing redundant egress rules", "err", err)
 		}
 
 		*lastKnownGoodEndpoint = "" // reset endpoint
-		if g.config.EnableServicesElection || g.config.EnableLeaderElection {
-			cancel() // stop services watcher
-		}
-		*leaderElectionActive = false
 	}
 }
 
-func (g *generic) getEndpoints(_ *v1.Service, id string) ([]string, error) {
-	return g.getLocalEndpoints(id)
-}
-
-func (g *generic) getLocalEndpoints(id string) ([]string, error) {
-	// Build endpoints
-	var endpoints []string
-	var err error
-	if endpoints, err = g.provider.GetLocalEndpoints(id, g.config); err != nil {
-		return nil, fmt.Errorf("[%s] error getting local endpoints: %w", g.provider.GetLabel(), err)
-	}
-
-	return endpoints, nil
+func (g *generic) getEndpoints(service *v1.Service, id string) ([]string, error) {
+	return g.getAllEndpoints(service, id)
 }
 
 func (g *generic) getAllEndpoints(service *v1.Service, id string) ([]string, error) {
 	// Build endpoints
 	var err error
 	var endpoints []string
-	if !g.config.EnableLeaderElection && !g.config.EnableServicesElection &&
-		service.Spec.ExternalTrafficPolicy == v1.ServiceExternalTrafficPolicyTypeCluster {
+	if service.Spec.ExternalTrafficPolicy == v1.ServiceExternalTrafficPolicyTypeCluster {
 		if endpoints, err = g.provider.GetAllEndpoints(); err != nil {
 			return nil, fmt.Errorf("[%s] error getting all endpoints: %w", g.provider.GetLabel(), err)
 		}
@@ -109,10 +105,10 @@ func (g *generic) getAllEndpoints(service *v1.Service, id string) ([]string, err
 func (g *generic) removeEgress(_ *v1.Service, _ *string) {
 }
 
-func (g *generic) delete(_ *v1.Service, _ string) error {
+func (g *generic) delete(_ context.Context, _ *v1.Service, _ string) error {
 	return nil
 }
 
-func (g *generic) setInstanceEndpointsStatus(_ *v1.Service, _ []string) error {
+func (g *generic) setInstanceEndpointsStatus(_ context.Context, _ *v1.Service, _ []string) error {
 	return nil
 }

pkg/endpoints/endpoints_routing_table.go

@@ -2,64 +2,46 @@ package endpoints
 
 import (
 	"context"
-	"errors"
 	"fmt"
 	"net"
-	"syscall"
+	"sync"
 
 	log "log/slog"
 
 	"github.com/kube-vip/kube-vip/pkg/egress"
 	"github.com/kube-vip/kube-vip/pkg/instance"
+	"github.com/kube-vip/kube-vip/pkg/lease"
+	"github.com/kube-vip/kube-vip/pkg/route"
 	"github.com/kube-vip/kube-vip/pkg/servicecontext"
-	"github.com/vishvananda/netlink"
 	v1 "k8s.io/api/core/v1"
 )
 
 type RoutingTable struct {
 	generic
+	mtx      sync.Mutex
+	routeMgr *route.Manager
 }
 
-func newRoutingTable(generic generic) endpointWorker {
+func newRoutingTable(generic generic, routeMgr *route.Manager) endpointWorker {
 	return &RoutingTable{
-		generic: generic,
+		generic:  generic,
+		routeMgr: routeMgr,
 	}
 }
 
-func (rt *RoutingTable) processInstance(ctx *servicecontext.Context, service *v1.Service, leaderElectionActive *bool) error {
-	instance := instance.FindServiceInstance(service, *rt.instances)
-	if instance != nil {
-		for _, cluster := range instance.Clusters {
+func (rt *RoutingTable) processInstance(svcCtx *servicecontext.Context, service *v1.Service) error {
+	inst := instance.FindServiceInstance(service, *rt.instances)
+	if inst != nil {
+		for _, cluster := range inst.Clusters {
 			for i := range cluster.Network {
-				if !ctx.IsNetworkConfigured(cluster.Network[i].IP()) && cluster.Network[i].HasEndpoints() {
-					err := cluster.Network[i].AddRoute(false)
-					if err != nil {
-						if errors.Is(err, syscall.EEXIST) {
-							// If route exists, but protocol is not set (e.g. the route was created by the older version
-							// of kube-vip) try to update it if necessary
-							isUpdated, err := cluster.Network[i].UpdateRoutes()
-							if err != nil {
-								return fmt.Errorf("[%s] error updating existing routes: %w", rt.provider.GetLabel(), err)
-							}
-							if isUpdated {
-								log.Info("updated route", "provider",
-									rt.provider.GetLabel(), "ip", cluster.Network[i].IP(), "service name", service.Name, "namespace",
-									service.Namespace, "interface", cluster.Network[i].Interface(), "tableID", rt.config.RoutingTableID)
-							} else {
-								log.Info("route already present", "provider",
-									rt.provider.GetLabel(), "ip", cluster.Network[i].IP(), "service name", service.Name, "namespace",
-									service.Namespace, "interface", cluster.Network[i].Interface(), "tableID", rt.config.RoutingTableID)
-							}
-						} else {
-							// If other error occurs, return error
-							return fmt.Errorf("[%s] error adding route: %s", rt.provider.GetLabel(), err.Error())
-						}
+				if !svcCtx.IsNetworkConfigured(cluster.Network[i].IP()) && cluster.Network[i].HasEndpoints() {
+					if err := rt.routeMgr.Add(lease.ServiceNamespacedName(service), cluster.Network[i], false, true); err != nil {
+						return fmt.Errorf("[%s] error adding route: %s", rt.provider.GetLabel(), err.Error())
 					} else {
 						log.Info("added route", "provider",
 							rt.provider.GetLabel(), "ip", cluster.Network[i].IP(), "service name", service.Name, "namespace",
 							service.Namespace, "interface", cluster.Network[i].Interface(), "tableID", rt.config.RoutingTableID)
-						ctx.ConfiguredNetworks.Store(cluster.Network[i].IP(), true)
-						*leaderElectionActive = true
+						svcCtx.ConfiguredNetworks.Store(cluster.Network[i].IP(), true)
 					}
 				}
 			}
@@ -69,9 +51,11 @@ func (rt *RoutingTable) processInstance(ctx *servicecontext.Context, service *v1
 	return nil
 }
 
-func (rt *RoutingTable) clear(svcCtx *servicecontext.Context, lastKnownGoodEndpoint *string, service *v1.Service, cancel context.CancelFunc, leaderElectionActive *bool) {
-	if !rt.config.EnableServicesElection && !rt.config.EnableLeaderElection {
-		if errs := ClearRoutes(service, rt.instances); len(errs) == 0 {
+func (rt *RoutingTable) clear(svcCtx *servicecontext.Context, lastKnownGoodEndpoint *string, service *v1.Service) {
+	rt.mtx.Lock()
+	defer rt.mtx.Unlock()
+	if !rt.config.EnableServicesElection {
+		if errs := ClearRoutes(service, rt.instances, rt.routeMgr); len(errs) == 0 {
 			svcCtx.ConfiguredNetworks.Clear()
 		} else {
 			for _, err := range errs {
@@ -80,7 +64,11 @@ func (rt *RoutingTable) clear(svcCtx *servicecontext.Context, lastKnownGoodEndpo
 		}
 	}
 
-	rt.clearEgress(lastKnownGoodEndpoint, service, cancel, leaderElectionActive)
+	rt.clearEgress(lastKnownGoodEndpoint, service)
+
+	if svcCtx.LeaderCancel != nil {
+		svcCtx.LeaderCancel()
+	}
 }
 
 func (rt *RoutingTable) getEndpoints(service *v1.Service, id string) ([]string, error) {
@@ -94,7 +82,7 @@ func (rt *RoutingTable) removeEgress(service *v1.Service, lastKnownGoodEndpoint
 	}
 }
 
-func (rt *RoutingTable) delete(service *v1.Service, id string) error {
+func (rt *RoutingTable) delete(_ context.Context, service *v1.Service, id string) error {
 	// When no-leader-elecition mode
 	if !rt.config.EnableServicesElection && !rt.config.EnableLeaderElection {
 		// find all existing local endpoints
@@ -113,23 +101,27 @@ func (rt *RoutingTable) delete(service *v1.Service, id string) error {
 }
 
 func (rt *RoutingTable) deleteAction(service *v1.Service) {
-	ClearRoutes(service, rt.instances)
+	ClearRoutes(service, rt.instances, rt.routeMgr)
 }
 
-func (rt *RoutingTable) setInstanceEndpointsStatus(service *v1.Service, endpoints []string) error {
-	instance := instance.FindServiceInstance(service, *rt.instances)
-	if instance == nil {
-		log.Error("failed to find the instance", "service", service.UID, "provider", rt.provider.GetLabel())
+func (rt *RoutingTable) setInstanceEndpointsStatus(ctx context.Context, service *v1.Service, endpoints []string) error {
+	inst := instance.FindServiceInstance(service, *rt.instances)
+	if inst == nil {
+		log.Error("failed to find the instance", "namespace", service.Namespace, "name", service.Name, "uid", service.UID, "provider", rt.provider.GetLabel())
 	} else {
-		for _, c := range instance.Clusters {
+		for _, c := range inst.Clusters {
 			for n := range c.Network {
 				// if there are no endpoints set HasEndpoints false just in case
 				if len(endpoints) < 1 {
 					c.Network[n].SetHasEndpoints(false)
-				}
-				// check if endpoint are available and are of same IP family as service
-				if len(endpoints) > 0 && ((net.ParseIP(c.Network[n].IP()).To4() == nil) == (net.ParseIP(endpoints[0]).To4() == nil)) {
-					c.Network[n].SetHasEndpoints(true)
+				} else {
+					// check if endpoint are available and are of same IP family as service
+					for _, ep := range endpoints {
+						if (net.ParseIP(c.Network[n].IP()).To4() == nil) == (net.ParseIP(ep).To4() == nil) {
+							c.Network[n].SetHasEndpoints(true)
+							break
+						}
+					}
 				}
 			}
 		}
@@ -138,41 +130,32 @@ func (rt *RoutingTable) setInstanceEndpointsStatus(service *v1.Service, endpoint
 	return nil
 }
 
-func ClearRoutes(service *v1.Service, instances *[]*instance.Instance) []error {
+func ClearRoutes(service *v1.Service, instances *[]*instance.Instance, routeMgr *route.Manager) []error {
 	errs := []error{}
-	if instance := instance.FindServiceInstance(service, *instances); instance != nil {
-		for _, cluster := range instance.Clusters {
-			for i := range cluster.Network {
-				route := cluster.Network[i].PrepareRoute()
-				// check if route we are about to delete is not referenced by more than one service
-				if CountRouteReferences(route, instances) <= 1 {
-					err := cluster.Network[i].DeleteRoute()
-					if err != nil && !errors.Is(err, syscall.ESRCH) {
-						log.Error("failed to delete route", "ip", cluster.Network[i].IP(), "err", err)
-						errs = append(errs, err)
-					}
-					log.Debug("deleted route", "ip",
-						cluster.Network[i].IP(), "service name", service.Name, "namespace", service.Namespace, "interface", cluster.Network[i].Interface())
-				}
-			}
-		}
+	if svcInst := instance.FindServiceInstance(service, *instances); svcInst != nil {
+		clearErrs := ClearRoutesByInstance(service, svcInst, instances, routeMgr)
+		errs = append(errs, clearErrs...)
 	}
 	return errs
 }
 
-func CountRouteReferences(route *netlink.Route, instances *[]*instance.Instance) int {
-	cnt := 0
-	for _, instance := range *instances {
-		for _, cluster := range instance.Clusters {
-			for n := range cluster.Network {
-				if cluster.Network[n].HasEndpoints() {
-					r := cluster.Network[n].PrepareRoute()
-					if r.Dst.String() == route.Dst.String() {
-						cnt++
-					}
-				}
+func ClearRoutesByInstance(service *v1.Service, svcInst *instance.Instance, instances *[]*instance.Instance, routeMgr *route.Manager) []error {
+	if svcInst == nil {
+		return []error{fmt.Errorf("failed to remove routes for nil instance of service %s/%s, uid: %s", service.Namespace, service.Name, service.UID)}
+	}
+	errs := []error{}
+	for _, cluster := range svcInst.Clusters {
+		for i := range cluster.Network {
+			err := routeMgr.Delete(lease.ServiceNamespacedName(service), cluster.Network[i])
+			if err != nil {
+				log.Error("failed to delete route", "ip", cluster.Network[i].IP(), "err", err)
+				errs = append(errs, err)
 			}
+			log.Debug("deleted route", "ip",
+				cluster.Network[i].IP(), "service name", service.Name, "namespace", service.Namespace, "interface", cluster.Network[i].Interface())
+
 		}
 	}
-	return cnt
+
+	return errs
 }

pkg/endpoints/endpoints_test.go

@@ -0,0 +1,405 @@
+package endpoints
+
+import (
+	"context"
+	"sync"
+	"sync/atomic"
+	"testing"
+	"time"
+
+	"github.com/kube-vip/kube-vip/pkg/endpoints/providers"
+	"github.com/kube-vip/kube-vip/pkg/instance"
+	"github.com/kube-vip/kube-vip/pkg/kubevip"
+	"github.com/kube-vip/kube-vip/pkg/lease"
+	"github.com/kube-vip/kube-vip/pkg/metrics"
+	"github.com/kube-vip/kube-vip/pkg/servicecontext"
+	"github.com/prometheus/client_golang/prometheus/testutil"
+	v1 "k8s.io/api/core/v1"
+	discoveryv1 "k8s.io/api/discovery/v1"
+	metav1 "k8s.io/apimachinery/pkg/apis/meta/v1"
+	"k8s.io/apimachinery/pkg/watch"
+	"k8s.io/client-go/kubernetes"
+)
+
+func TestShouldAllowReconcileWithoutEndpoints(t *testing.T) {
+	if shouldAllowReconcileWithoutEndpoints(nil) {
+		t.Fatal("nil service should not be allowed")
+	}
+
+	clusterOptIn := &v1.Service{
+		Spec:       v1.ServiceSpec{ExternalTrafficPolicy: v1.ServiceExternalTrafficPolicyTypeCluster},
+		ObjectMeta: metav1.ObjectMeta{Annotations: map[string]string{kubevip.AllowReconcileWithoutEndpoints: "true"}},
+	}
+	if !shouldAllowReconcileWithoutEndpoints(clusterOptIn) {
+		t.Fatal("cluster service with opt-in annotation should be allowed")
+	}
+
+	localOptIn := &v1.Service{
+		Spec:       v1.ServiceSpec{ExternalTrafficPolicy: v1.ServiceExternalTrafficPolicyTypeLocal},
+		ObjectMeta: metav1.ObjectMeta{Annotations: map[string]string{kubevip.AllowReconcileWithoutEndpoints: "true"}},
+	}
+	if shouldAllowReconcileWithoutEndpoints(localOptIn) {
+		t.Fatal("local service should not be allowed")
+	}
+}
+
+type fakeWorker struct {
+	endpoints     []string
+	clearCalled   bool
+	processCalled bool
+}
+
+type annotationUpdate struct {
+	endpoint     string
+	endpointIPv6 string
+}
+
+type recordingProvider struct {
+	providers.Provider
+	updates []annotationUpdate
+}
+
+func (p *recordingProvider) UpdateServiceAnnotation(_ context.Context, endpoint, endpointIPv6 string,
+	_ *v1.Service, _ *kubernetes.Clientset) error {
+	p.updates = append(p.updates, annotationUpdate{endpoint: endpoint, endpointIPv6: endpointIPv6})
+	return nil
+}
+
+func TestUpdateAnnotationsZeroEndpointsThenSameEndpoint(t *testing.T) {
+	for _, enableEndpoints := range []bool{true, false} {
+		providerName := "EndpointSlices"
+		provider := providers.NewEndpointslices()
+		if enableEndpoints {
+			providerName = "Endpoints"
+			provider = providers.NewEndpoints()
+		}
+
+		for _, family := range []struct {
+			name       string
+			endpoint   string
+			other      string
+			egressIPv6 bool
+		}{
+			{name: "IPv4", endpoint: "10.0.0.1", other: "fd00::1"},
+			{name: "IPv6", endpoint: "fd00::1", other: "10.0.0.1", egressIPv6: true},
+		} {
+			t.Run(providerName+"/"+family.name, func(t *testing.T) {
+				annotations := map[string]string{kubevip.Egress: "true"}
+				if family.egressIPv6 {
+					annotations[kubevip.EgressIPv6] = "true"
+				}
+				if !enableEndpoints {
+					if family.egressIPv6 {
+						annotations[kubevip.ActiveEndpoint] = family.other
+						annotations[kubevip.ActiveEndpointIPv6] = family.endpoint
+					} else {
+						annotations[kubevip.ActiveEndpoint] = family.endpoint
+						annotations[kubevip.ActiveEndpointIPv6] = family.other
+					}
+				} else {
+					annotations[kubevip.ActiveEndpoint] = family.endpoint
+				}
+				service := &v1.Service{ObjectMeta: metav1.ObjectMeta{
+					Name: "test-service", Namespace: "default", UID: "test-uid", Annotations: annotations,
+				}}
+				serviceInstance := &instance.Instance{ServiceSnapshot: service.DeepCopy()}
+				instances := []*instance.Instance{serviceInstance}
+				recorder := &recordingProvider{Provider: provider}
+				processor := &Processor{
+					config:    &kubevip.Config{EnableEndpoints: enableEndpoints},
+					provider:  recorder,
+					instances: &instances,
+				}
+
+				updateSnapshot := func(_ context.Context, updated *v1.Service) error {
+					serviceInstance.ServiceSnapshot = updated
+					return nil
+				}
+
+				noEndpoint := ""
+				processor.updateAnnotations(service, &noEndpoint, nil, updateSnapshot)
+				repopulatedEndpoint := family.endpoint
+				processor.updateAnnotations(service, &repopulatedEndpoint, nil, updateSnapshot)
+
+				cleared := annotationUpdate{}
+				repopulated := annotationUpdate{endpoint: family.endpoint}
+				if !enableEndpoints {
+					if family.egressIPv6 {
+						cleared = annotationUpdate{endpoint: family.other}
+						repopulated = annotationUpdate{endpoint: family.other, endpointIPv6: family.endpoint}
+					} else {
+						cleared = annotationUpdate{endpointIPv6: family.other}
+						repopulated = annotationUpdate{endpoint: family.endpoint, endpointIPv6: family.other}
+					}
+				}
+				want := []annotationUpdate{cleared, repopulated}
+				if len(recorder.updates) != len(want) {
+					t.Fatalf("annotation updates = %+v, want %+v", recorder.updates, want)
+				}
+				for index := range want {
+					if recorder.updates[index] != want[index] {
+						t.Errorf("annotation update %d = %+v, want %+v", index, recorder.updates[index], want[index])
+					}
+				}
+			})
+		}
+	}
+}
+
+func TestUpdateAnnotationsEndpointSlicesClearsConfiguredFamily(t *testing.T) {
+	for _, test := range []struct {
+		name       string
+		egressIPv6 bool
+		want       annotationUpdate
+	}{
+		{name: "IPv4", want: annotationUpdate{endpointIPv6: "fd00::1"}},
+		{name: "IPv6", egressIPv6: true, want: annotationUpdate{endpoint: "10.0.0.1"}},
+	} {
+		t.Run(test.name, func(t *testing.T) {
+			annotations := map[string]string{
+				kubevip.Egress:             "true",
+				kubevip.ActiveEndpoint:     "10.0.0.1",
+				kubevip.ActiveEndpointIPv6: "fd00::1",
+			}
+			if test.egressIPv6 {
+				annotations[kubevip.EgressIPv6] = "true"
+			}
+			service := &v1.Service{ObjectMeta: metav1.ObjectMeta{
+				Name: "test-service", Namespace: "default", UID: "test-uid", Annotations: annotations,
+			}}
+			instances := []*instance.Instance{{ServiceSnapshot: service.DeepCopy()}}
+			recorder := &recordingProvider{Provider: providers.NewEndpointslices()}
+			processor := &Processor{
+				config:    &kubevip.Config{EnableEndpoints: false},
+				provider:  recorder,
+				instances: &instances,
+			}
+
+			noEndpoint := ""
+			processor.updateAnnotations(service, &noEndpoint, nil, func(context.Context, *v1.Service) error { return nil })
+
+			if len(recorder.updates) != 1 || recorder.updates[0] != test.want {
+				t.Fatalf("annotation updates = %+v, want [%+v]", recorder.updates, test.want)
+			}
+		})
+	}
+}
+
+func TestUpdateAnnotationsValidatesEndpointFamily(t *testing.T) {
+	tests := []struct {
+		name       string
+		endpoint   string
+		egressIPv6 bool
+		want       annotationUpdate
+		wantUpdate bool
+	}{
+		{name: "invalid address", endpoint: "not-an-ip"},
+		{name: "IPv6 endpoint for IPv4 egress", endpoint: "fd00::1"},
+		{name: "IPv4 endpoint for IPv6 egress", endpoint: "10.0.0.1", egressIPv6: true},
+		{name: "IPv4 endpoint", endpoint: "10.0.0.2", want: annotationUpdate{endpoint: "10.0.0.2", endpointIPv6: "fd00::1"}, wantUpdate: true},
+		{name: "IPv6 endpoint", endpoint: "fd00::2", egressIPv6: true, want: annotationUpdate{endpoint: "10.0.0.1", endpointIPv6: "fd00::2"}, wantUpdate: true},
+	}
+
+	for _, test := range tests {
+		t.Run(test.name, func(t *testing.T) {
+			annotations := map[string]string{
+				kubevip.Egress:             "true",
+				kubevip.ActiveEndpoint:     "10.0.0.1",
+				kubevip.ActiveEndpointIPv6: "fd00::1",
+			}
+			if test.egressIPv6 {
+				annotations[kubevip.EgressIPv6] = "true"
+			}
+			service := &v1.Service{ObjectMeta: metav1.ObjectMeta{
+				Name: "test-service", Namespace: "default", Annotations: annotations,
+			}}
+			recorder := &recordingProvider{Provider: providers.NewEndpointslices()}
+			processor := &Processor{
+				config:   &kubevip.Config{EnableEndpoints: false},
+				provider: recorder,
+			}
+
+			processor.updateAnnotations(service, &test.endpoint, nil, nil)
+
+			if !test.wantUpdate {
+				if len(recorder.updates) != 0 {
+					t.Fatalf("annotation updates = %+v, want none", recorder.updates)
+				}
+				return
+			}
+			if len(recorder.updates) != 1 || recorder.updates[0] != test.want {
+				t.Fatalf("annotation updates = %+v, want [%+v]", recorder.updates, test.want)
+			}
+		})
+	}
+}
+
+func (f *fakeWorker) processInstance(_ *servicecontext.Context, _ *v1.Service) error {
+	f.processCalled = true
+	return nil
+}
+
+func (f *fakeWorker) clear(_ *servicecontext.Context, _ *string, _ *v1.Service) {
+	f.clearCalled = true
+}
+
+func (f *fakeWorker) getEndpoints(_ *v1.Service, _ string) ([]string, error) { return f.endpoints, nil }
+func (f *fakeWorker) removeEgress(_ *v1.Service, _ *string)                  {}
+func (f *fakeWorker) delete(_ context.Context, _ *v1.Service, _ string) error {
+	return nil
+}
+func (f *fakeWorker) setInstanceEndpointsStatus(_ context.Context, _ *v1.Service, _ []string) error {
+	return nil
+}
+
+func TestAddOrModify_ZeroEndpointsBehavior(t *testing.T) {
+	t.Parallel()
+
+	run := func(t *testing.T, service *v1.Service, presetSignalled bool, expectReady bool, expectClear bool, expectProcess bool) {
+		t.Helper()
+
+		worker := &fakeWorker{endpoints: []string{}}
+		p := &Processor{
+			config:   &kubevip.Config{},
+			provider: providers.NewEndpointslices(),
+			worker:   worker,
+		}
+
+		svcCtx := servicecontext.New(context.Background())
+		if presetSignalled {
+			svcCtx.SignalReadiness()
+		}
+
+		restart, err := p.AddOrModify(
+			svcCtx,
+			watch.Event{Type: watch.Modified, Object: &discoveryv1.EndpointSlice{}},
+			new(string),
+			service,
+			"node-1",
+			func(*servicecontext.Context, *v1.Service, *sync.WaitGroup, bool) error { return nil },
+			&sync.WaitGroup{},
+			nil,
+			nil,
+		)
+		if err != nil {
+			t.Fatalf("AddOrModify returned error: %v", err)
+		}
+		if restart {
+			t.Fatal("AddOrModify unexpectedly requested restart")
+		}
+
+		if ready := svcCtx.Signalled.Load(); ready != expectReady {
+			t.Fatalf("readiness mismatch: expected %v, got %v", expectReady, ready)
+		}
+		if worker.clearCalled != expectClear {
+			t.Fatalf("clearCalled mismatch: expected %v, got %v", expectClear, worker.clearCalled)
+		}
+		if worker.processCalled != expectProcess {
+			t.Fatalf("processCalled mismatch: expected %v, got %v", expectProcess, worker.processCalled)
+		}
+	}
+
+	t.Run("cluster opt-in keeps readiness and skips clear", func(t *testing.T) {
+		service := &v1.Service{
+			ObjectMeta: metav1.ObjectMeta{Annotations: map[string]string{kubevip.AllowReconcileWithoutEndpoints: "true"}},
+			Spec:       v1.ServiceSpec{ExternalTrafficPolicy: v1.ServiceExternalTrafficPolicyTypeCluster},
+		}
+		run(t, service, false, true, false, true)
+	})
+
+	t.Run("cluster without opt-in resets and clears when pre-signalled", func(t *testing.T) {
+		service := &v1.Service{
+			ObjectMeta: metav1.ObjectMeta{Annotations: map[string]string{}},
+			Spec:       v1.ServiceSpec{ExternalTrafficPolicy: v1.ServiceExternalTrafficPolicyTypeCluster},
+		}
+		run(t, service, true, false, true, false)
+	})
+
+	t.Run("local opt-in still resets and clears when pre-signalled", func(t *testing.T) {
+		service := &v1.Service{
+			ObjectMeta: metav1.ObjectMeta{Annotations: map[string]string{kubevip.AllowReconcileWithoutEndpoints: "true"}},
+			Spec:       v1.ServiceSpec{ExternalTrafficPolicy: v1.ServiceExternalTrafficPolicyTypeLocal},
+		}
+		run(t, service, true, false, true, false)
+	})
+}
+
+// TestAddOrModify_ServicesElectionStartsOnce asserts that repeated endpoint events
+// for the same service start the leader-election restart loop exactly once.
+//
+// AddOrModify runs on every EndpointSlice add/modify/resync event, and the loop it
+// starts only returns once the service context is cancelled. Starting it per event
+// therefore accumulates duplicate goroutines that all contend on the same lease.
+//
+// See https://github.com/kube-vip/kube-vip/issues/1665.
+func TestAddOrModify_ServicesElectionStartsOnce(t *testing.T) {
+	config := &kubevip.Config{
+		EnableServicesElection: true,
+		LeaderElectionType:     "kubernetes",
+	}
+
+	service := &v1.Service{
+		ObjectMeta: metav1.ObjectMeta{Name: "test-svc", Namespace: "default", UID: "test-uid"},
+		Spec:       v1.ServiceSpec{Type: v1.ServiceTypeLoadBalancer},
+	}
+
+	ctx, cancel := context.WithCancel(context.Background())
+	defer cancel()
+
+	leaseMgr := lease.NewManager()
+	leaseNamespace, serviceLease := lease.ServiceName(service)
+	svcLease := leaseMgr.Add(ctx, lease.NewID(config.LeaderElectionType, leaseNamespace, serviceLease))
+
+	svcCtx := servicecontext.New(svcLease.Ctx)
+
+	// The started loops only return once the service context is cancelled, so it has
+	// to be cancelled before waiting on them.
+	wg := &sync.WaitGroup{}
+	defer wg.Wait()
+	defer svcCtx.Cancel()
+
+	p := &Processor{
+		config:   config,
+		provider: providers.NewEndpointslices(),
+		worker:   &fakeWorker{endpoints: []string{"10.0.0.1"}},
+		leaseMgr: leaseMgr,
+	}
+
+	// starts counts the restart loops. The real StartServicesLeaderElection blocks
+	// until the service context is cancelled, so each loop parks in a single call.
+	var starts atomic.Int64
+	serviceFunc := func(svcCtx *servicecontext.Context, _ *v1.Service, _ *sync.WaitGroup, _ bool) error {
+		starts.Add(1)
+		<-svcCtx.Ctx.Done()
+		return nil
+	}
+
+	// Three endpoint events, as a flapping backend pod would produce.
+	for range 3 {
+		restart, err := p.AddOrModify(svcCtx, watch.Event{Type: watch.Modified, Object: &discoveryv1.EndpointSlice{}},
+			new(string), service, "node-1", serviceFunc, wg, nil, nil)
+		if err != nil {
+			t.Fatalf("AddOrModify returned error: %v", err)
+		}
+		if restart {
+			t.Fatal("AddOrModify unexpectedly requested restart")
+		}
+	}
+
+	// Give every loop that is going to start a chance to reach serviceFunc.
+	for deadline := time.Now().Add(2 * time.Second); time.Now().Before(deadline); {
+		if starts.Load() > 1 {
+			break
+		}
+		time.Sleep(10 * time.Millisecond)
+	}
+
+	if got := starts.Load(); got != 1 {
+		t.Errorf("leader election started %d times, want 1", got)
+	}
+
+	// The gauge the e2e fault tests assert on has to agree with the call count.
+	if got := testutil.ToFloat64(metrics.ServiceElectionLoops.WithLabelValues(service.Namespace, service.Name)); got != 1 {
+		t.Errorf("kube_vip_service_election_loops is %v, want 1", got)
+	}
+}

pkg/endpoints/endpoints_wireguard.go

@@ -0,0 +1,289 @@
+package endpoints
+
+import (
+	"context"
+	"fmt"
+
+	log "log/slog"
+
+	"github.com/kube-vip/kube-vip/pkg/bgp"
+	"github.com/kube-vip/kube-vip/pkg/endpoints/providers"
+	"github.com/kube-vip/kube-vip/pkg/instance"
+	"github.com/kube-vip/kube-vip/pkg/kubevip"
+	"github.com/kube-vip/kube-vip/pkg/lease"
+	"github.com/kube-vip/kube-vip/pkg/nftables"
+	"github.com/kube-vip/kube-vip/pkg/servicecontext"
+	"github.com/kube-vip/kube-vip/pkg/utils"
+	"github.com/kube-vip/kube-vip/pkg/wireguard"
+	v1 "k8s.io/api/core/v1"
+)
+
+// wireguardWorker handles endpoint changes for WireGuard-based services
+type wireguardWorker struct {
+	config    *kubevip.Config
+	provider  providers.Provider
+	bgpServer *bgp.Server
+	instances *[]*instance.Instance
+	leaseMgr  *lease.Manager
+	tunnelMgr *wireguard.TunnelManager
+}
+
+func newWireguardWorker(config *kubevip.Config, provider providers.Provider, bgpServer *bgp.Server,
+	instances *[]*instance.Instance, leaseMgr *lease.Manager, tunnelMgr *wireguard.TunnelManager) *wireguardWorker {
+	return &wireguardWorker{
+		config:    config,
+		provider:  provider,
+		bgpServer: bgpServer,
+		instances: instances,
+		leaseMgr:  leaseMgr,
+		tunnelMgr: tunnelMgr,
+	}
+}
+
+// processInstance updates nftables DNAT rules when endpoints change
+// This is called by the endpoint watcher when endpoints are added/modified
+func (w *wireguardWorker) processInstance(svcCtx *servicecontext.Context, service *v1.Service) error {
+	log.Debug("[wireguard] processing instance for endpoint change", "service", service.Name, "namespace", service.Namespace)
+
+	// Get the target endpoint for this service
+	// For ExternalTrafficPolicy=Local, only use local endpoints
+	// For ExternalTrafficPolicy=Cluster, use all endpoints
+	var endpoints []string
+	var err error
+	if service.Spec.ExternalTrafficPolicy == v1.ServiceExternalTrafficPolicyTypeLocal {
+		endpoints, err = w.provider.GetLocalEndpoints(w.config.NodeName, w.config)
+	} else {
+		endpoints, err = w.provider.GetAllEndpoints()
+	}
+	if err != nil {
+		return fmt.Errorf("failed to get endpoints: %w", err)
+	}
+
+	if len(endpoints) == 0 {
+		log.Debug("[wireguard] no endpoints available", "service", service.Name)
+		w.clear(svcCtx, nil, service)
+		return nil
+	}
+
+	// Find the service processor to call updateServiceWireguardEndpoints
+	// Note: This requires access to the service processor which we don't have here
+	// So we'll recreate the DNAT rules directly
+
+	// First, clear existing rules
+	w.clear(svcCtx, nil, service)
+
+	// Get service VIPs
+	serviceIPs, err := utils.FetchServiceIPs(service)
+	if err != nil {
+		return fmt.Errorf("failed to get service IPs: %w", err)
+	}
+
+	// Create service identifier
+	serviceID := utils.SanitizeServiceID(fmt.Sprintf("%s_%s", service.Namespace, service.Name))
+
+	log.Info("[wireguard] updating DNAT rules for endpoint change",
+		"service", service.Name,
+		"namespace", service.Namespace,
+		"endpoints", endpoints,
+		"vips", serviceIPs)
+
+	// Update DNAT rules for each port
+	for _, port := range service.Spec.Ports {
+		// Determine target port (resolve named ports if necessary)
+		targetPort := w.provider.ResolvePort(port)
+		log.Info("[wireguard] resolved port", "service", service.Name, "servicePort", port.Port, "targetPort", targetPort, "targetPortName", port.TargetPort.StrVal)
+
+		// Build targets list from all endpoints
+		targets := make([]nftables.DNATTarget, len(endpoints))
+		for i, ep := range endpoints {
+			targets[i] = nftables.DNATTarget{
+				IP:   ep,
+				Port: uint16(targetPort), //nolint:gosec // Port range validated by Kubernetes
+			}
+		}
+
+		for _, vip := range serviceIPs {
+			// Strip CIDR notation if present
+			vipAddr := utils.StripCIDR(vip)
+
+			// Get WireGuard interface name from TunnelManager for this VIP
+			if w.tunnelMgr == nil {
+				log.Error("[wireguard] TunnelManager not configured; cannot update DNAT rules",
+					"service", service.Name,
+					"namespace", service.Namespace)
+				return fmt.Errorf("TunnelManager not configured")
+			}
+			tunnelConfig := w.tunnelMgr.GetConfigForVIP(vipAddr)
+			if tunnelConfig == nil {
+				log.Error("[wireguard] WireGuard interface name not configured; cannot update DNAT rules",
+					"service", service.Name,
+					"namespace", service.Namespace,
+					"vip", vipAddr)
+				return fmt.Errorf("wireguard interface name not configured for VIP %s", vipAddr)
+			}
+			wgInterface := tunnelConfig.InterfaceName
+
+			portServiceID := fmt.Sprintf("%s_p%d", serviceID, port.Port)
+
+			log.Info("[wireguard] applying DNAT rule with load balancing",
+				"service", service.Name,
+				"vip", vipAddr,
+				"interface", wgInterface,
+				"sourcePort", port.Port,
+				"targets", targets,
+				"chainID", portServiceID)
+
+			// Apply the DNAT rule with load balancing across all endpoints
+			// localEndpoint=true when using ExternalTrafficPolicy=Local, which preserves client source IP
+			isLocalEndpoint := service.Spec.ExternalTrafficPolicy == v1.ServiceExternalTrafficPolicyTypeLocal
+			err := nftables.ApplyDNAT(
+				wgInterface,
+				vipAddr,
+				uint16(port.Port), //nolint:gosec // Port range validated by Kubernetes
+				targets,
+				portServiceID,
+				port.Protocol,
+				isLocalEndpoint,
+				tunnelConfig.ListenPort,
+			)
+			if err != nil {
+				log.Error("[wireguard] failed to update DNAT rule",
+					"service", service.Name,
+					"vip", vipAddr,
+					"port", port.Port,
+					"err", err)
+				continue
+			}
+
+			log.Debug("[wireguard] DNAT rule updated successfully",
+				"service", service.Name,
+				"vip", vipAddr,
+				"port", port.Port,
+				"targetCount", len(targets))
+		}
+	}
+
+	return nil
+}
+
+// clear removes DNAT rules when no endpoints are available
+func (w *wireguardWorker) clear(svcCtx *servicecontext.Context, lastKnownGoodEndpoint *string, service *v1.Service) {
+	log.Info("[wireguard] clearing DNAT rules (no endpoints)", "service", service.Name, "namespace", service.Namespace)
+
+	serviceID := utils.SanitizeServiceID(fmt.Sprintf("%s_%s", service.Namespace, service.Name))
+
+	// Get service IPs to determine IPv4 vs IPv6
+	serviceIPs, _ := utils.FetchServiceIPs(service)
+
+	// Delete DNAT chains for each port
+	for _, port := range service.Spec.Ports {
+		if port.Protocol != v1.ProtocolTCP && port.Protocol != v1.ProtocolUDP {
+			continue
+		}
+
+		portServiceID := fmt.Sprintf("%s_p%d", serviceID, port.Port)
+
+		// Determine if we have IPv4 or IPv6
+		hasIPv4, hasIPv6 := false, false
+		for _, vip := range serviceIPs {
+			if isIPv6Address(vip) {
+				hasIPv6 = true
+			} else {
+				hasIPv4 = true
+			}
+		}
+
+		if hasIPv4 {
+			if err := nftables.DeleteIngressChains(false, portServiceID); err != nil {
+				log.Warn("[wireguard] failed to delete IPv4 DNAT chains",
+					"service", service.Name,
+					"port", port.Port,
+					"err", err)
+			}
+		}
+
+		if hasIPv6 {
+			if err := nftables.DeleteIngressChains(true, portServiceID); err != nil {
+				log.Warn("[wireguard] failed to delete IPv6 DNAT chains",
+					"service", service.Name,
+					"port", port.Port,
+					"err", err)
+			}
+		}
+	}
+
+	if svcCtx.LeaderCancel != nil {
+		svcCtx.LeaderCancel()
+	}
+}
+
+// getEndpoints retrieves the list of endpoints for a service
+// For ExternalTrafficPolicy=Local, only local endpoints are returned
+// For ExternalTrafficPolicy=Cluster, all endpoints are returned
+func (w *wireguardWorker) getEndpoints(service *v1.Service, id string) ([]string, error) {
+	var endpoints []string
+	var err error
+	if service.Spec.ExternalTrafficPolicy == v1.ServiceExternalTrafficPolicyTypeLocal {
+		endpoints, err = w.provider.GetLocalEndpoints(id, w.config)
+	} else {
+		endpoints, err = w.provider.GetAllEndpoints()
+	}
+	if err != nil {
+		return nil, fmt.Errorf("[wireguard] failed to get endpoints: %w", err)
+	}
+
+	log.Debug("[wireguard] retrieved endpoints", "service", service.Name, "count", len(endpoints), "endpoints", endpoints)
+	return endpoints, nil
+}
+
+// removeEgress is a no-op for WireGuard since egress is handled separately
+func (w *wireguardWorker) removeEgress(service *v1.Service, lastKnownGoodEndpoint *string) {
+	// WireGuard doesn't use egress in the same way as other modes
+	log.Debug("[wireguard] removeEgress called (no-op)", "service", service.Name)
+}
+
+// delete removes all DNAT rules for a service
+func (w *wireguardWorker) delete(ctx context.Context, service *v1.Service, id string) error {
+	log.Info("[wireguard] deleting DNAT rules for service", "service", service.Name, "namespace", service.Namespace)
+
+	w.clear(nil, nil, service)
+	return nil
+}
+
+// setInstanceEndpointsStatus updates the endpoint status on the service instance
+func (w *wireguardWorker) setInstanceEndpointsStatus(_ context.Context, service *v1.Service, endpoints []string) error {
+	hasEndpoints := len(endpoints) > 0
+
+	log.Debug("[wireguard] setting instance endpoint status",
+		"service", service.Name,
+		"hasEndpoints", hasEndpoints,
+		"endpointCount", len(endpoints))
+
+	// Find the service instance
+	for _, inst := range *w.instances {
+		if inst.ServiceSnapshot == nil {
+			continue
+		}
+		if inst.ServiceSnapshot.UID == service.UID {
+			// Update the network status for all clusters
+			for _, cluster := range inst.Clusters {
+				for i := range cluster.Network {
+					cluster.Network[i].SetHasEndpoints(hasEndpoints)
+				}
+			}
+			log.Debug("[wireguard] updated instance endpoint status",
+				"service", service.Name,
+				"hasEndpoints", hasEndpoints)
+			return nil
+		}
+	}
+
+	log.Debug("[wireguard] instance not found for endpoint status update", "service", service.Name)
+	return nil
+}
+
+func isIPv6Address(ip string) bool {
+	// Strip CIDR notation if present before checking
+	addr := utils.StripCIDR(ip)
+	return utils.IsIPv6(addr)
+}

pkg/endpoints/providers/endpoints.go

@@ -97,11 +97,11 @@ func (ep *Endpoints) GetLocalEndpoints(id string, _ *kubevip.Config) ([]string,
 	return localEndpoints, nil
 }
 
-func (ep *Endpoints) UpdateServiceAnnotation(endpoint string, _ string, service *v1.Service, clientSet *kubernetes.Clientset) error {
+func (ep *Endpoints) UpdateServiceAnnotation(ctx context.Context, endpoint string, _ string, service *v1.Service, clientSet *kubernetes.Clientset) error {
 	retryErr := retry.RetryOnConflict(retry.DefaultRetry, func() error {
 		// Retrieve the latest version of Deployment before attempting update
 		// RetryOnConflict uses exponential backoff to avoid exhausting the apiserver
-		currentService, err := clientSet.CoreV1().Services(service.Namespace).Get(context.TODO(), service.Name, metav1.GetOptions{})
+		currentService, err := clientSet.CoreV1().Services(service.Namespace).Get(ctx, service.Name, metav1.GetOptions{})
 		if err != nil {
 			return err
 		}
@@ -113,7 +113,7 @@ func (ep *Endpoints) UpdateServiceAnnotation(endpoint string, _ string, service
 
 		currentServiceCopy.Annotations[kubevip.ActiveEndpoint] = endpoint
 
-		_, err = clientSet.CoreV1().Services(currentService.Namespace).Update(context.TODO(), currentServiceCopy, metav1.UpdateOptions{})
+		_, err = clientSet.CoreV1().Services(currentService.Namespace).Update(ctx, currentServiceCopy, metav1.UpdateOptions{})
 		if err != nil {
 			log.Error("error updating Service Spec", "label", ep.GetLabel(), "name", currentServiceCopy.Name, "err", err)
 			return err
@@ -132,6 +132,15 @@ func (ep *Endpoints) GetLabel() string {
 	return ep.label
 }
 
-func (ep *Endpoints) GetProtocol() string {
-	return ""
+func (ep *Endpoints) ResolvePort(servicePort v1.ServicePort) int32 {
+	return ResolvePortWithLookup(servicePort, func(name string) int32 {
+		for _, subset := range ep.endpoints.Subsets {
+			for _, p := range subset.Ports {
+				if p.Name == name {
+					return p.Port
+				}
+			}
+		}
+		return 0
+	})
 }

pkg/endpoints/providers/endpointslices.go

@@ -20,8 +20,10 @@ import (
 )
 
 type Endpointslices struct {
-	label     string
-	endpoints *discoveryv1.EndpointSlice
+	label       string
+	endpointsv4 []discoveryv1.Endpoint
+	endpointsv6 []discoveryv1.Endpoint
+	ports       []discoveryv1.EndpointPort
 }
 
 func NewEndpointslices() Provider {
@@ -56,27 +58,42 @@ func (ep *Endpointslices) LoadObject(endpoints runtime.Object, cancel context.Ca
 		cancel()
 		return fmt.Errorf("[%s] error casting endpoints to v1.Endpoints struct", ep.label)
 	}
-	ep.endpoints = eps
+
+	if eps.AddressType == discoveryv1.AddressTypeIPv6 {
+		ep.endpointsv6 = eps.Endpoints
+	} else {
+		ep.endpointsv4 = eps.Endpoints
+	}
+
+	// Store ports for resolving named ports
+	ep.ports = eps.Ports
+
 	return nil
 }
 
 func (ep *Endpointslices) GetAllEndpoints() ([]string, error) {
 	result := []string{}
-	for _, ep := range ep.endpoints.Endpoints {
-		result = append(result, ep.Addresses...)
+	for _, e := range ep.endpointsv4 {
+		result = append(result, e.Addresses...)
+	}
+	for _, e := range ep.endpointsv6 {
+		result = append(result, e.Addresses...)
 	}
 	return result, nil
 }
 
 func (ep *Endpointslices) GetLocalEndpoints(id string, _ *kubevip.Config) ([]string, error) {
 	var localEndpoints []string
-	for _, endpoint := range ep.endpoints.Endpoints {
-		if !*endpoint.Conditions.Serving {
+	tmpEps := []discoveryv1.Endpoint{}
+
+	tmpEps = append(tmpEps, ep.endpointsv4...)
+	tmpEps = append(tmpEps, ep.endpointsv6...)
+
+	for _, endpoint := range tmpEps {
+		if endpoint.Conditions.Serving == nil || !*endpoint.Conditions.Serving {
 			continue
 		}
 		for _, address := range endpoint.Addresses {
-			log.Debug("processing endpoint", "provider", ep.label, "ip", address)
-
 			// 1. Compare the Nodename
 			if endpoint.NodeName != nil && id == *endpoint.NodeName {
 				if endpoint.Hostname != nil {
@@ -98,11 +115,11 @@ func (ep *Endpointslices) GetLocalEndpoints(id string, _ *kubevip.Config) ([]str
 	return localEndpoints, nil
 }
 
-func (ep *Endpointslices) UpdateServiceAnnotation(endpoint, endpointIPv6 string, service *v1.Service, clientSet *kubernetes.Clientset) error {
+func (ep *Endpointslices) UpdateServiceAnnotation(ctx context.Context, endpoint, endpointIPv6 string, service *v1.Service, clientSet *kubernetes.Clientset) error {
 	retryErr := retry.RetryOnConflict(retry.DefaultRetry, func() error {
 		// Retrieve the latest version of Deployment before attempting update
 		// RetryOnConflict uses exponential backoff to avoid exhausting the apiserver
-		currentService, err := clientSet.CoreV1().Services(service.Namespace).Get(context.TODO(), service.Name, metav1.GetOptions{})
+		currentService, err := clientSet.CoreV1().Services(service.Namespace).Get(ctx, service.Name, metav1.GetOptions{})
 		if err != nil {
 			return err
 		}
@@ -115,7 +132,7 @@ func (ep *Endpointslices) UpdateServiceAnnotation(endpoint, endpointIPv6 string,
 		currentServiceCopy.Annotations[kubevip.ActiveEndpoint] = endpoint
 		currentServiceCopy.Annotations[kubevip.ActiveEndpointIPv6] = endpointIPv6
 
-		_, err = clientSet.CoreV1().Services(currentService.Namespace).Update(context.TODO(), currentServiceCopy, metav1.UpdateOptions{})
+		_, err = clientSet.CoreV1().Services(currentService.Namespace).Update(ctx, currentServiceCopy, metav1.UpdateOptions{})
 		if err != nil {
 			log.Error("error updating Service Spec", "provider", ep.label, "service name", currentServiceCopy.Name, "err", err)
 			return err
@@ -134,6 +151,13 @@ func (ep *Endpointslices) GetLabel() string {
 	return ep.label
 }
 
-func (ep *Endpointslices) GetProtocol() string {
-	return string(ep.endpoints.AddressType)
+func (ep *Endpointslices) ResolvePort(servicePort v1.ServicePort) int32 {
+	return ResolvePortWithLookup(servicePort, func(name string) int32 {
+		for _, p := range ep.ports {
+			if p.Name != nil && *p.Name == name && p.Port != nil {
+				return *p.Port
+			}
+		}
+		return 0
+	})
 }

pkg/endpoints/providers/interface.go

@@ -16,7 +16,24 @@ type Provider interface {
 	GetAllEndpoints() ([]string, error)
 	GetLocalEndpoints(string, *kubevip.Config) ([]string, error)
 	GetLabel() string
-	UpdateServiceAnnotation(string, string, *v1.Service, *kubernetes.Clientset) error
+	UpdateServiceAnnotation(context.Context, string, string, *v1.Service, *kubernetes.Clientset) error
 	LoadObject(runtime.Object, context.CancelFunc) error
-	GetProtocol() string
+	// ResolvePort resolves a service port to the actual target port.
+	// For named ports, it looks up the port number from the endpoint.
+	// For numeric ports, it returns the port as-is.
+	ResolvePort(servicePort v1.ServicePort) int32
+}
+
+// ResolvePortWithLookup is a helper that resolves a service port using a lookup function
+// for named ports. This consolidates the common resolution logic.
+func ResolvePortWithLookup(servicePort v1.ServicePort, lookupNamedPort func(string) int32) int32 {
+	if servicePort.TargetPort.IntVal != 0 {
+		return servicePort.TargetPort.IntVal
+	}
+	if servicePort.TargetPort.StrVal != "" {
+		if port := lookupNamedPort(servicePort.TargetPort.StrVal); port != 0 {
+			return port
+		}
+	}
+	return servicePort.Port
 }

pkg/etcd/election.go

@@ -2,7 +2,9 @@ package etcd
 
 import (
 	"context"
+	"fmt"
 	"hash/fnv"
+	"sync"
 	"time"
 
 	log "log/slog"
@@ -64,10 +66,11 @@ type ClientConfig struct {
 }
 
 // RunElectionOrDie behaves the same way as RunElection but panics if there is an error.
-func RunElectionOrDie(ctx context.Context, config *LeaderElectionConfig) {
+func RunElectionOrDie(ctx context.Context, config *LeaderElectionConfig) error {
 	if err := RunElection(ctx, config); err != nil {
-		panic(err)
+		return fmt.Errorf("leaderelection error: %w", err)
 	}
+	return nil
 }
 
 // RunElection starts a client with the provided config or panics.
@@ -116,7 +119,12 @@ func RunElection(ctx context.Context, config *LeaderElectionConfig) error {
 		leaseTTL:       lease.TTL,
 	}
 
-	go m.tryToBeLeader(ctx)
+	wg := sync.WaitGroup{}
+	defer wg.Wait()
+
+	wg.Go(func() {
+		m.tryToBeLeader(ctx, &wg)
+	})
 	m.watchLeaderChanges(ctx)
 
 	return nil
@@ -185,7 +193,7 @@ watcher:
 	log.Debug("Exiting watcher", "id", m.memberID)
 }
 
-func (m *member) tryToBeLeader(ctx context.Context) {
+func (m *member) tryToBeLeader(ctx context.Context, wg *sync.WaitGroup) {
 	if err := m.election.Campaign(ctx, m.memberID); err != nil {
 		log.Error("Failed trying to become the leader", "err", err)
 		// Resign just in case we acquired leadership just before failing
@@ -204,7 +212,9 @@ func (m *member) tryToBeLeader(ctx context.Context) {
 	m.weAreTheLeader <- struct{}{}
 
 	// Once we are the leader, start the routine to resign if context is canceled
-	go m.resignOnCancel(ctx)
+	wg.Go(func() {
+		m.resignOnCancel(ctx)
+	})
 
 	// After becoming the leader, we wait for at least a lease TTL to wait for
 	// the previous leader to detect the new leadership (if there was one) and
@@ -221,7 +231,7 @@ func (m *member) tryToBeLeader(ctx context.Context) {
 
 func (m *member) resignOnCancel(ctx context.Context) {
 	<-ctx.Done()
-	if err := m.election.Resign(m.client.Ctx()); err != nil {
+	if err := m.election.Resign(m.client.Ctx()); err != nil && !errors.Is(err, context.Canceled) {
 		log.Error("Failed to resign after the context was canceled", "err", err)
 	}
 }

pkg/etcd/election_test.go

@@ -20,19 +20,26 @@ import (
 func TestRunElectionWithMemberIDCollision(t *testing.T) {
 	t.Parallel()
 	g := NewWithT(t)
-	ctx := context.Background()
+	ctx, cancel := context.WithCancel(context.Background())
+	defer cancel()
 	cli := client(g)
 	defer cli.Close()
 
 	electionName := randomElectionNameForTest("memberIDConflict")
 	log.Printf("Election name %s\n", electionName)
 	memberCtx, cancelMember1 := context.WithCancel(ctx)
+
+	// Use a channel to signal when the first member has observed a new leader
+	// This ensures proper ordering without relying on sleep timing
+	firstMemberObservedLeader := make(chan struct{})
+	var firstMemberObservedOnce sync.Once
+
 	config := &etcd.LeaderElectionConfig{
 		EtcdConfig: etcd.ClientConfig{
 			Client: cli,
 		},
 		Name:                 electionName,
-		MemberID:             "my-host",
+		MemberID:             randomElectionNameForTest("my-host"),
 		LeaseDurationSeconds: 1,
 		Callbacks: etcd.LeaderCallbacks{
 			OnStartedLeading: func(ctx context.Context) {
@@ -42,6 +49,11 @@ func TestRunElectionWithMemberIDCollision(t *testing.T) {
 			},
 			OnNewLeader: func(identity string) {
 				log.Printf("New leader: %s\n", identity)
+				// Signal that the first member has observed a leader
+				// This means the lease has been created
+				firstMemberObservedOnce.Do(func() {
+					close(firstMemberObservedLeader)
+				})
 			},
 			OnStoppedLeading: func() {
 				log.Println("I'm not the leader anymore")
@@ -59,8 +71,18 @@ func TestRunElectionWithMemberIDCollision(t *testing.T) {
 
 	go func() {
 		defer wg.Done()
-		time.Sleep(time.Millisecond * 50) // make sure the first one becomes leader
-		g.Expect(etcd.RunElection(ctx, config)).Should(MatchError(ContainSubstring("creating lease")))
+		// Wait for the first member to observe a leader, which means the lease has been created
+		select {
+		case <-firstMemberObservedLeader:
+			// First member has created the lease, now try to create a conflicting one
+		case <-time.After(5 * time.Second):
+			t.Error("timeout waiting for first member to observe leader")
+			return
+		}
+		// Use a cancellable context to prevent hanging if this goroutine unexpectedly succeeds
+		member2Ctx, cancelMember2 := context.WithTimeout(ctx, 5*time.Second)
+		defer cancelMember2()
+		g.Expect(etcd.RunElection(member2Ctx, config)).Should(MatchError(ContainSubstring("creating lease")))
 	}()
 
 	wg.Wait()
@@ -69,7 +91,8 @@ func TestRunElectionWithMemberIDCollision(t *testing.T) {
 func TestRunElectionWithTwoMembersAndReelection(t *testing.T) {
 	t.Parallel()
 	g := NewWithT(t)
-	ctx := context.Background()
+	ctx, cancel := context.WithCancel(context.Background())
+	defer cancel()
 	cli := client(g)
 	defer cli.Close()
 
@@ -85,24 +108,27 @@ func TestRunElectionWithTwoMembersAndReelection(t *testing.T) {
 		LeaseDurationSeconds: 1,
 	}
 
-	member1Ctx, _ := context.WithCancel(ctx)
+	member1Ctx, cancelMember1 := context.WithCancel(ctx)
 	member2Ctx, cancelMember2 := context.WithCancel(ctx)
 
 	config1 := configBase
 	config1.EtcdConfig.Client = cliMember1
-	config1.MemberID = "my-host"
+	config1.MemberID = randomElectionNameForTest("my-host")
 	uniqueID := rand.Uint64()
 	config1.MemberUniqueID = &uniqueID
 	config1.Callbacks = baseCallbacksForName(config1.MemberID)
+	syncMembers := make(chan (any))
 	config1.Callbacks.OnStartedLeading = func(_ context.Context) {
 		log.Println("I'm my-host, the new leader!!!!")
+		close(syncMembers)
 		log.Println("Losing the leadership on purpose by stopping renewing the lease")
 		g.Expect(cliMember1.Lease.Close()).To(Succeed())
 		log.Println("Member1 leases closed")
+		cancelMember1()
 	}
 
 	config2 := configBase
-	config2.MemberID = "my-other-host"
+	config2.MemberID = randomElectionNameForTest("my-other-host")
 	config2.Callbacks = baseCallbacksForName(config2.MemberID)
 	config2.Callbacks.OnStartedLeading = func(_ context.Context) {
 		log.Println("I'm my-other-host, the new leader!!!!")
@@ -116,17 +142,18 @@ func TestRunElectionWithTwoMembersAndReelection(t *testing.T) {
 	go func() {
 		defer wg.Done()
 		g.Expect(etcd.RunElection(member1Ctx, &config1)).To(Succeed())
-		log.Println("Member1 routine done")
+		log.Printf("%s routine done\n", config1.MemberID)
 	}()
 
 	go func() {
 		defer wg.Done()
-		time.Sleep(time.Millisecond * 50) // Make sure member1 becomes leader
+		<-syncMembers
 		g.Expect(etcd.RunElection(member2Ctx, &config2)).To(Succeed())
-		log.Println("Member2 routine done")
+		log.Printf("%s routine done\n", config2.MemberID)
 	}()
 
 	wg.Wait()
+
 }
 
 func baseCallbacksForName(name string) etcd.LeaderCallbacks {

pkg/etcd/etcd_suite_test.go

@@ -15,7 +15,6 @@ import (
 	"time"
 
 	"github.com/pkg/errors"
-	"github.com/sirupsen/logrus"
 )
 
 const (
@@ -28,8 +27,8 @@ const (
 )
 
 func TestMain(m *testing.M) {
-	logrus.SetLevel(logrus.DebugLevel)
-	ctx := context.Background()
+	ctx, cancel := context.WithCancel(context.Background())
+	defer cancel()
 	expectSuccess(startEtcd(ctx), "starting etcd")
 
 	os.Exit(runTestsWithCleanup(m, func() {

pkg/instance/instance.go

@@ -1,12 +1,14 @@
 package instance
 
 import (
+	"context"
 	"fmt"
 	"net"
+	"slices"
 	"strconv"
 	"strings"
+	"sync"
 
-	"log/slog"
 	log "log/slog"
 
 	"github.com/vishvananda/netlink"
@@ -16,7 +18,10 @@ import (
 	"github.com/kube-vip/kube-vip/pkg/cluster"
 	"github.com/kube-vip/kube-vip/pkg/kubevip"
 	"github.com/kube-vip/kube-vip/pkg/networkinterface"
+	"github.com/kube-vip/kube-vip/pkg/node"
+	"github.com/kube-vip/kube-vip/pkg/route"
 	"github.com/kube-vip/kube-vip/pkg/sysctl"
+	"github.com/kube-vip/kube-vip/pkg/utils"
 	"github.com/kube-vip/kube-vip/pkg/vip"
 )
 
@@ -29,18 +34,37 @@ type Instance struct {
 	Clusters []*cluster.Cluster
 
 	// Service uses DHCP
-	IsDHCP              bool
+	IsDHCPv4            bool
+	IsDHCPv6            bool
 	DHCPInterface       string
 	DHCPInterfaceHwaddr string
 	DHCPInterfaceIP     string
+	DHCPInterfaceIPv4   string
+	DHCPInterfaceIPv6   string
 	DHCPHostname        string
-	DHCPClient          *vip.DHCPClient
+	DHCPv4Client        vip.DHCPClient
+	DHCPv6Client        vip.DHCPClient
+	macvlanName         string
+	dhcpBroadcast       bool
+
+	// Service use Vlan
+	IsVLAN        bool
+	VLANInterface string
 
 	// External Gateway IP the service is forwarded from
 	UPNPGatewayIPs []string
 
 	// Kubernetes service mapping
 	ServiceSnapshot *v1.Service
+
+	dnsAddresses []string
+
+	// AddCalled determined that ActionAdd was already performed for the instance
+	AddCalled bool
+
+	// LabelAdded determined that node was labeled with
+	// service-provided.kube-vip.io label
+	LabelAdded bool
 }
 
 type Port struct {
@@ -48,18 +72,42 @@ type Port struct {
 	Type string
 }
 
-func NewInstance(svc *v1.Service, config *kubevip.Config, intfMgr *networkinterface.Manager, arpMgr *arp.Manager) (*Instance, error) {
-	instanceAddresses := FetchServiceAddresses(svc)
-	//instanceUID := string(svc.UID)
+func NewInstance(ctx context.Context, svc *v1.Service, config *kubevip.Config,
+	intfMgr *networkinterface.Manager, arpMgr *arp.Manager, routeMgr *route.Manager,
+	nodeLabelMgr node.Labeler, wg *sync.WaitGroup) (*Instance, error) {
+	instanceAddresses, instanceHostnames := FetchServiceAddresses(svc)
+	log.Info("new instance", "namespace", svc.Namespace, "service", svc.Name, "addresses", instanceAddresses, "hostnames", instanceHostnames)
 
 	var newVips []*kubevip.Config
 	var link netlink.Link
 	var err error
+	var dnsAddresses []string
+
+	// Create new service
+	instance := &Instance{
+		ServiceSnapshot: svc,
+		dnsAddresses:    dnsAddresses,
+	}
 
 	for _, address := range instanceAddresses {
 		// Detect if we're using a specific interface for services
 		var svcInterface string
-		svcInterface = svc.Annotations[kubevip.ServiceInterface] // If the service has a specific interface defined, then use it
+
+		svcInterface = svc.Annotations[kubevip.ServiceVlan]
+		if svcInterface != "" {
+			parent, tag, err := utils.ParseVLANInterface(svcInterface)
+			if err != nil {
+				log.Error("failed to validate VLAN", "err", err)
+			}
+
+			if err := instance.addVLAN(parent, tag); err != nil {
+				log.Error("failed to create VLAN", "err", err)
+			}
+		} else {
+			// If no vlan defined use specific interface from annotation
+			svcInterface = svc.Annotations[kubevip.ServiceInterface]
+		}
+
 		if svcInterface == kubevip.Auto {
 			link, err = autoFindInterface(address)
 			if err != nil {
@@ -120,7 +168,7 @@ func NewInstance(svc *v1.Service, config *kubevip.Config, intfMgr *networkinterf
 
 		subnet := ""
 		var err error
-		if vip.IsIPv4(address) {
+		if utils.IsIPv4(address) {
 			if ipv4AutoSubnet {
 				subnet, err = autoFindSubnet(link, address)
 				if err != nil {
@@ -130,7 +178,7 @@ func NewInstance(svc *v1.Service, config *kubevip.Config, intfMgr *networkinterf
 				if cidrs[0] != "" && cidrs[0] != kubevip.Auto {
 					subnet = cidrs[0]
 				} else {
-					subnet = "32"
+					subnet = strconv.Itoa(vip.DefaultMaskIPv4)
 				}
 			}
 		} else {
@@ -143,26 +191,94 @@ func NewInstance(svc *v1.Service, config *kubevip.Config, intfMgr *networkinterf
 				if len(cidrs) > 1 && cidrs[1] != "" && cidrs[1] != kubevip.Auto {
 					subnet = cidrs[1]
 				} else {
-					subnet = "128"
+					subnet = strconv.Itoa(vip.DefaultMaskIPv6)
 				}
 			}
 		}
 
 		// Generate new Virtual IP configuration
 		newVips = append(newVips, &kubevip.Config{
-			VIP:                    address,
+			VIP:                         address,
+			Interface:                   svcInterface,
+			SingleNode:                  true,
+			EnableARP:                   config.EnableARP,
+			EnableBGP:                   config.EnableBGP,
+			VIPSubnet:                   subnet,
+			EnableRoutingTable:          config.EnableRoutingTable,
+			RoutingTableID:              config.RoutingTableID,
+			RoutingTableType:            config.RoutingTableType,
+			RoutingProtocol:             config.RoutingProtocol,
+			SkipDAD:                     config.SkipDAD,
+			ArpBroadcastRate:            config.ArpBroadcastRate,
+			EnableServiceSecurity:       config.EnableServiceSecurity,
+			DNSMode:                     config.DNSMode,
+			DHCPMode:                    config.DHCPMode,
+			DHCPBackoffAttempts:         config.DHCPBackoffAttempts,
+			DisableServiceUpdates:       config.DisableServiceUpdates,
+			EnableServicesElection:      config.EnableServicesElection,
+			PreserveVIPOnLeadershipLoss: config.PreserveVIPOnLeadershipLoss,
+			KubernetesLeaderElection: kubevip.KubernetesLeaderElection{
+				EnableLeaderElection: config.EnableLeaderElection,
+			},
+		})
+	}
+
+	for _, hostname := range instanceHostnames {
+		log.Info("hostname", "addr", hostname)
+		// Detect if we're using a specific interface for services
+		var svcInterface string
+
+		svcInterface = svc.Annotations[kubevip.ServiceVlan]
+		if svcInterface != "" {
+			parent, tag, err := utils.ParseVLANInterface(svcInterface)
+			if err != nil {
+				log.Error("failed to validate VLAN", "err", err)
+			}
+
+			if err := instance.addVLAN(parent, tag); err != nil {
+				log.Error("failed to create VLAN", "err", err)
+			}
+		} else {
+			// If no vlan defined use specific interface from annotation
+			svcInterface = svc.Annotations[kubevip.ServiceInterface]
+		}
+
+		// If it is still blank then use the
+		if svcInterface == "" {
+			switch config.ServicesInterface {
+			case "":
+				svcInterface = config.Interface
+			default:
+				svcInterface = config.ServicesInterface
+			}
+		}
+
+		if link == nil {
+			if link, err = netlink.LinkByName(svcInterface); err != nil {
+				return nil, fmt.Errorf("failed to get interface %s: %w", svcInterface, err)
+			}
+			if link == nil {
+				return nil, fmt.Errorf("failed to get interface %s", svcInterface)
+			}
+		}
+
+		// Generate new Virtual IP configuration
+		newVips = append(newVips, &kubevip.Config{
+			VIP:                    hostname,
 			Interface:              svcInterface,
 			SingleNode:             true,
 			EnableARP:              config.EnableARP,
 			EnableBGP:              config.EnableBGP,
-			VIPSubnet:              subnet,
+			VIPSubnet:              config.VIPSubnet,
 			EnableRoutingTable:     config.EnableRoutingTable,
 			RoutingTableID:         config.RoutingTableID,
 			RoutingTableType:       config.RoutingTableType,
 			RoutingProtocol:        config.RoutingProtocol,
+			SkipDAD:                config.SkipDAD,
 			ArpBroadcastRate:       config.ArpBroadcastRate,
 			EnableServiceSecurity:  config.EnableServiceSecurity,
 			DNSMode:                config.DNSMode,
+			DHCPMode:               config.DHCPMode,
 			DisableServiceUpdates:  config.DisableServiceUpdates,
 			EnableServicesElection: config.EnableServicesElection,
 			KubernetesLeaderElection: kubevip.KubernetesLeaderElection{
@@ -171,23 +287,26 @@ func NewInstance(svc *v1.Service, config *kubevip.Config, intfMgr *networkinterf
 		})
 	}
 
-	// Create new service
-	instance := &Instance{
-		//UID:             instanceUID,
-		//VIPs:            instanceAddresses,
-		ServiceSnapshot: svc,
-	}
-	// for _, port := range svc.Spec.Ports {
-	// 	instance.ExternalPorts = append(instance.ExternalPorts, Port{
-	// 		Port: uint16(port.Port), //nolint
-	// 		Type: string(port.Protocol),
-	// 	})
-	// }
-
 	if svc.Annotations != nil {
 		instance.DHCPInterfaceHwaddr = svc.Annotations[kubevip.HwAddrKey]
-		instance.DHCPInterfaceIP = svc.Annotations[kubevip.RequestedIP]
+		requestedIP := svc.Annotations[kubevip.RequestedIP]
+		if requestedIP != "" {
+			requestedIPs := strings.Split(requestedIP, ",")
+			if len(requestedIPs) > 2 {
+				return nil, fmt.Errorf("annotation %q cannot request more than one IPv4 and one Ipv6 address", kubevip.RequestedIP)
+			}
+			for _, ip := range requestedIPs {
+				netip := net.ParseIP(ip)
+				if netip.To4() != nil {
+					instance.DHCPInterfaceIPv4 = ip
+				} else {
+					instance.DHCPInterfaceIPv6 = ip
+				}
+			}
+		}
 		instance.DHCPHostname = svc.Annotations[kubevip.LoadbalancerHostname]
+		instance.macvlanName = svc.Annotations[kubevip.MacvlanName]
+		instance.dhcpBroadcast = svc.Annotations[kubevip.DHCPBroadcast] == "true"
 	}
 
 	configPorts := make([]kubevip.Port, 0)
@@ -210,29 +329,84 @@ func NewInstance(svc *v1.Service, config *kubevip.Config, intfMgr *networkinterf
 	// Create Add configuration to the new service
 	instance.VIPConfigs = newVips
 
-	// If this was purposely created with the address 0.0.0.0,
+	// If this was purposely created with the address '0.0.0.0', or '::'
 	// we will create a macvlan on the main interface and a DHCP client
-	// TODO: Consider how best to handle DHCP with multiple addresses
-	if len(instanceAddresses) == 1 && instanceAddresses[0] == "0.0.0.0" {
-		err := instance.startDHCP()
-		if err != nil {
-			return nil, err
-		}
-		select {
-		case err := <-instance.DHCPClient.ErrorChannel():
-			return nil, fmt.Errorf("error starting DHCP for %s/%s: error: %s",
-				instance.ServiceSnapshot.Namespace, instance.ServiceSnapshot.Name, err)
-		case ip := <-instance.DHCPClient.IPChannel():
-			instance.VIPConfigs[0].Interface = instance.DHCPInterface
-			instance.VIPConfigs[0].VIP = ip
-			instance.DHCPInterfaceIP = ip
-		}
+	if len(instanceAddresses) > 2 && (slices.Contains(instanceAddresses, "0.0.0.0") || slices.Contains(instanceAddresses, "::")) {
+		return nil, fmt.Errorf("DHCP cannot be used if more than 2 addresses (one IPv4 and one IPv6) were specified")
 	}
+	for i := range instance.VIPConfigs {
+		if instance.VIPConfigs[i].VIP == "0.0.0.0" {
+			err := instance.startDHCP(ctx, i, config.DHCPBackoffAttempts, wg)
+			if err != nil {
+				return nil, err
+			}
+			select {
+			case err := <-instance.DHCPv4Client.ErrorChannel():
+				return nil, fmt.Errorf("error starting DHCPv4 for %s/%s: error: %s",
+					instance.ServiceSnapshot.Namespace, instance.ServiceSnapshot.Name, err)
+			case ip := <-instance.DHCPv4Client.IPChannel():
+				instance.VIPConfigs[i].Interface = instance.DHCPInterface
+				instance.VIPConfigs[i].VIP = ip
+				instance.DHCPInterfaceIPv4 = ip
+			}
+		}
+		if instance.VIPConfigs[i].VIP == "::" {
+			err := instance.startDHCP(ctx, i, config.DHCPBackoffAttempts, wg)
+			if err != nil {
+				return nil, err
+			}
+			select {
+			case err := <-instance.DHCPv6Client.ErrorChannel():
+				return nil, fmt.Errorf("error starting DHCPv6 for %s/%s: error: %s",
+					instance.ServiceSnapshot.Namespace, instance.ServiceSnapshot.Name, err)
+			case ip := <-instance.DHCPv6Client.IPChannel():
+				instance.VIPConfigs[i].Interface = instance.DHCPInterface
+				instance.VIPConfigs[i].VIP = ip
+				instance.DHCPInterfaceIPv6 = ip
+			}
+		}
 
-	for _, vipConfig := range instance.VIPConfigs {
-		c, err := cluster.InitCluster(vipConfig, false, intfMgr, arpMgr)
+		ddnsAnnotation, exists := svc.Annotations[kubevip.ServiceDDNS]
+
+		if exists {
+			instance.VIPConfigs[i].DDNS, err = strconv.ParseBool(ddnsAnnotation)
+			if err != nil {
+				log.Error("Failed to add service", "err", err)
+				return nil, err
+			}
+		}
+
+		if len(svc.Spec.IPFamilies) > 0 {
+			if len(svc.Spec.IPFamilies) > 1 {
+				instance.VIPConfigs[i].DHCPMode = utils.DualFamily
+				instance.VIPConfigs[i].DNSMode = utils.DualFamily
+				switch *svc.Spec.IPFamilyPolicy {
+				case v1.IPFamilyPolicyRequireDualStack:
+					instance.VIPConfigs[i].IsDualStack = true
+					instance.VIPConfigs[i].RequireDualStack = true
+				case v1.IPFamilyPolicyPreferDualStack:
+					instance.VIPConfigs[i].IsDualStack = true
+					instance.VIPConfigs[i].RequireDualStack = false
+				default:
+					instance.VIPConfigs[i].IsDualStack = false
+					instance.VIPConfigs[i].RequireDualStack = false
+				}
+			} else {
+				if strings.EqualFold(string(svc.Spec.IPFamilies[0]), utils.IPv4Family) {
+					instance.VIPConfigs[i].DHCPMode = strings.ToLower(utils.IPv4Family)
+					instance.VIPConfigs[i].DNSMode = strings.ToLower(utils.IPv4Family)
+				} else {
+					instance.VIPConfigs[i].DHCPMode = strings.ToLower(utils.IPv6Family)
+					instance.VIPConfigs[i].DNSMode = strings.ToLower(utils.IPv6Family)
+				}
+			}
+		}
+
+		instance.VIPConfigs[i].EgressWithNftables = config.EgressWithNftables
+
+		c, err := cluster.InitCluster(instance.VIPConfigs[i], false, intfMgr, arpMgr, routeMgr, nodeLabelMgr)
 		if err != nil {
-			log.Error("Failed to add Service %s/%s", svc.Namespace, svc.Name)
+			log.Error("failed to add service", "err", err)
 			return nil, err
 		}
 
@@ -241,7 +415,7 @@ func NewInstance(svc *v1.Service, config *kubevip.Config, intfMgr *networkinterf
 		}
 
 		instance.Clusters = append(instance.Clusters, c)
-		log.Info("(svcs) adding VIP", "ip", vipConfig.VIP, "interface", vipConfig.Interface, "namespace", svc.Namespace, "name", svc.Name)
+		log.Info("(svcs) adding VIP", "ip", instance.VIPConfigs[i].VIP, "interface", instance.VIPConfigs[i].Interface, "namespace", svc.Namespace, "name", svc.Name)
 	}
 
 	return instance, nil
@@ -304,17 +478,72 @@ func getAutoInterfaceName(link netlink.Link, defaultInterface string) string {
 	return link.Attrs().Name
 }
 
-func (i *Instance) startDHCP() error {
-	if len(i.VIPConfigs) != 1 {
-		return fmt.Errorf("DHCP requires exactly 1 VIP config, got: %v", len(i.VIPConfigs))
+func (i *Instance) addVLAN(parentInterface string, tag int) error {
+	var parent netlink.Link
+
+	interfaceName := fmt.Sprintf("%s.%d", parentInterface, tag)
+	iface, err := netlink.LinkByName(interfaceName)
+	if err != nil {
+		// check if parent interface doesnt exist
+		parent, err = netlink.LinkByName(parentInterface)
+		if err != nil {
+			return fmt.Errorf("error finding VLAN parent interface %s:  %v", parentInterface, err)
+		}
+
+		log.Info("Creating new VLAN interface", "interface", interfaceName)
+
+		vlan := &netlink.Vlan{
+			LinkAttrs: netlink.LinkAttrs{
+				Name:        interfaceName,
+				ParentIndex: parent.Attrs().Index,
+			},
+			VlanId:       tag,
+			VlanProtocol: netlink.VLAN_PROTOCOL_8021Q,
+		}
+
+		err = netlink.LinkAdd(vlan)
+		if err != nil {
+			return fmt.Errorf("could not add VLAN %s: %v", interfaceName, err)
+		}
+
+		err = netlink.LinkSetUp(vlan)
+		if err != nil {
+			return fmt.Errorf("could not bring up VLAN interface [%s] : %v", interfaceName, err)
+		}
+
+		_, err = net.InterfaceByName(interfaceName)
+		if err != nil {
+			return fmt.Errorf("error finding new VLAN interface by name [%v]", err)
+		}
+	} else {
+		log.Info("Using existing VLAN interface", "interface", interfaceName)
+
+		if err := utils.ValidateVLANInterface(iface, parent, tag); err != nil {
+			return err
+		}
 	}
-	parent, err := netlink.LinkByName(i.VIPConfigs[0].Interface)
+
+	i.VLANInterface = interfaceName
+	i.IsVLAN = true
+
+	return nil
+}
+
+func (i *Instance) startDHCP(ctx context.Context, index int, backoffAttempts uint, wg *sync.WaitGroup) error {
+	if len(i.VIPConfigs) > 2 {
+		return fmt.Errorf("DHCP can be used with 2 VIP config maximally, got: %v", len(i.VIPConfigs))
+	}
+	parent, err := netlink.LinkByName(i.VIPConfigs[index].Interface)
 	if err != nil {
 		return fmt.Errorf("error finding VIP Interface, for building DHCP Link : %v", err)
 	}
 
-	// Generate name from UID
-	interfaceName := fmt.Sprintf("vip-%s", i.ServiceSnapshot.UID[0:8])
+	interfaceName := i.macvlanName
+
+	if interfaceName == "" {
+		// Generate name from UID
+		interfaceName = fmt.Sprintf("vip-%s", i.ServiceSnapshot.UID[0:8])
+	}
 
 	// Check if the interface doesn't exist first
 	iface, err := net.InterfaceByName(interfaceName)
@@ -359,34 +588,61 @@ func (i *Instance) startDHCP() error {
 		log.Info("Using existing macvlan interface for DHCP", "interface", interfaceName)
 	}
 
-	// Default rp_filter setting (https://github.com/kube-vip/kube-vip/issues/1170)
-	rpfilterSetting := "0"
+	var initRebootFlag bool
+	ip := net.ParseIP(i.VIPConfigs[index].VIP)
 
-	// Check if we need to set an override rp_filter value for the interface
-	if i.ServiceSnapshot.Annotations[kubevip.RPFilter] != "" {
-		// Check the rp_filter value
-		rpFilter, err := strconv.Atoi(i.ServiceSnapshot.Annotations[kubevip.RPFilter])
-		if err != nil {
-			slog.Error("[DHCP] unable to process rp_filter", "value", rpFilter)
-		} else {
-			if rpFilter >= 0 && rpFilter < 3 { // Ensure the value is 0,1,2
-				rpfilterSetting = i.ServiceSnapshot.Annotations[kubevip.RPFilter]
+	var client vip.DHCPClient
+	if ip.To4() != nil {
+		// Default rp_filter setting (https://github.com/kube-vip/kube-vip/issues/1170)
+		rpfilterSetting := "0"
+
+		// Check if we need to set an override rp_filter value for the interface
+		if i.ServiceSnapshot.Annotations[kubevip.RPFilter] != "" {
+			// Check the rp_filter value
+			rpFilter, err := strconv.Atoi(i.ServiceSnapshot.Annotations[kubevip.RPFilter])
+			if err != nil {
+				log.Error("[DHCP] unable to process rp_filter", "value", rpFilter)
 			} else {
-				slog.Error("[DHCP] rp_filter value not within range 0-2", "value", rpFilter)
+				if rpFilter >= 0 && rpFilter < 3 { // Ensure the value is 0,1,2
+					rpfilterSetting = i.ServiceSnapshot.Annotations[kubevip.RPFilter]
+				} else {
+					log.Error("[DHCP] rp_filter value not within range 0-2", "value", rpFilter)
+				}
 			}
 		}
-	}
 
-	err = sysctl.WriteProcSys("/proc/sys/net/ipv4/conf/"+interfaceName+"/rp_filter", rpfilterSetting)
-	if err != nil {
-		slog.Error("[DHCP] unable to write rp_filter", "value", rpfilterSetting, "err", err)
-	}
-	var initRebootFlag bool
-	if i.DHCPInterfaceIP != "" {
-		initRebootFlag = true
-	}
+		err = sysctl.WriteProcSys("/proc/sys/net/ipv4/conf/"+interfaceName+"/rp_filter", rpfilterSetting)
+		if err != nil {
+			log.Error("[DHCP] unable to write rp_filter", "value", rpfilterSetting, "err", err)
+		}
 
-	client := vip.NewDHCPClient(iface, initRebootFlag, i.DHCPInterfaceIP)
+		if i.DHCPInterfaceIPv4 != "" {
+			initRebootFlag = true
+		}
+
+		client = vip.NewDHCPv4Client(iface, initRebootFlag, i.DHCPInterfaceIPv4, backoffAttempts, i.dhcpBroadcast)
+
+		// Add the client so that we can call it to stop function
+		i.DHCPv4Client = client
+
+		// Set that DHCPv4 is enabled
+		i.IsDHCPv4 = true
+	} else {
+		if i.DHCPInterfaceIPv6 != "" {
+			initRebootFlag = true
+		}
+
+		client, err = vip.NewDHCPv6Client(iface, parent, initRebootFlag, i.DHCPInterfaceIPv6, backoffAttempts)
+		if err != nil {
+			return fmt.Errorf("unable to create client: %w", err)
+		}
+
+		// Add the client so that we can call it to stop function
+		i.DHCPv6Client = client
+
+		// Set that DHCPv6 is enabled
+		i.IsDHCPv6 = true
+	}
 
 	// Add hostname to dhcp client if annotated
 	if i.DHCPHostname != "" {
@@ -394,89 +650,99 @@ func (i *Instance) startDHCP() error {
 		client.WithHostName(i.DHCPHostname)
 	}
 
-	go client.Start()
+	wg.Go(func() {
+		if err := client.Start(ctx); err != nil {
+			log.Error("[instance] DHCP client error: %w")
+		}
+	})
 
-	// Set that DHCP is enabled
-	i.IsDHCP = true
 	// Set the name of the interface so that it can be removed on Service deletion
 	i.DHCPInterface = interfaceName
 	i.DHCPInterfaceHwaddr = iface.HardwareAddr.String()
-	// Add the client so that we can call it to stop function
-	i.DHCPClient = client
 
 	return nil
 }
 
-// FetchIngressAddresses tries to get the addresses from status.loadBalancerIP
-func FetchLoadBalancerIngressAddresses(s *v1.Service) []string {
+// FetchLoadBalancerIngressAddresses tries to get the addresses from status.loadBalancerIP
+func FetchLoadBalancerIngress(s *v1.Service) ([]string, []string) {
 	// If the service has no status, return empty
 	lbStatusAddresses := []string{}
-
+	lbStatusHostnames := []string{}
 	if len(s.Status.LoadBalancer.Ingress) == 0 {
-		return lbStatusAddresses
+		return lbStatusAddresses, lbStatusHostnames
 	}
 
 	for _, ingress := range s.Status.LoadBalancer.Ingress {
 		if ingress.IP != "" {
 			lbStatusAddresses = append(lbStatusAddresses, ingress.IP)
 		}
-		// TODO: Handle hostname if needed
+		if ingress.Hostname != "" {
+			lbStatusHostnames = append(lbStatusHostnames, ingress.Hostname)
+		}
 	}
-
-	return lbStatusAddresses
+	return lbStatusAddresses, lbStatusHostnames
 }
 
 // FetchServiceAddresses tries to get the addresses from annotations
 // kube-vip.io/loadbalancerIPs, then from spec.loadbalancerIP
-func FetchServiceAddresses(s *v1.Service) []string {
+func FetchServiceAddresses(s *v1.Service) ([]string, []string) {
 	annotationAvailable := false
 	if s.Annotations != nil {
 
 		if v, annotationAvailable := s.Annotations[kubevip.LoadbalancerIPAnnotation]; annotationAvailable {
 			ips := strings.Split(v, ",")
 			var trimmedIPs []string
-			for _, ip := range ips {
-				trimmedIPs = append(trimmedIPs, strings.TrimSpace(ip))
+			var trimmedHostnames []string
+			for _, a := range ips {
+				a = strings.TrimSpace(a)
+				ip := net.ParseIP(a)
+				if ip == nil {
+					// this is probably a DNS name
+					trimmedHostnames = append(trimmedHostnames, a)
+				} else {
+					trimmedIPs = append(trimmedIPs, ip.String())
+				}
 			}
-			return trimmedIPs
+			return trimmedIPs, trimmedHostnames
 		}
 	}
 
 	lbStatusAddresses := []string{}
+	lbStatusHostnames := []string{}
 	if !annotationAvailable {
-		if len(s.Status.LoadBalancer.Ingress) > 0 {
-			for _, ingress := range s.Status.LoadBalancer.Ingress {
-				lbStatusAddresses = append(lbStatusAddresses, ingress.IP)
+		lbStatusAddresses, lbStatusHostnames = FetchLoadBalancerIngress(s)
+	}
+
+	// Spec.LoadBalancerIP legacy handling
+	// if the loadBalancerIP is different from Status.LoadBalancer.Ingress IPs
+	// return the legacy LB as spec wins over status.
+	if lbIP := net.ParseIP(s.Spec.LoadBalancerIP); lbIP != nil && len(lbStatusAddresses) > 0 {
+		isLbIPv4 := utils.IsIPv4(s.Spec.LoadBalancerIP)
+		for _, a := range lbStatusAddresses {
+			if lbStatusIP := net.ParseIP(a); lbStatusIP != nil && utils.IsIPv4(a) == isLbIPv4 && !lbIP.Equal(lbStatusIP) {
+				return []string{s.Spec.LoadBalancerIP}, []string{}
 			}
 		}
 	}
 
-	lbIP := net.ParseIP(s.Spec.LoadBalancerIP)
-	isLbIPv4 := vip.IsIPv4(s.Spec.LoadBalancerIP)
-
-	if len(lbStatusAddresses) > 0 {
-		for _, a := range lbStatusAddresses {
-			if lbStatusIP := net.ParseIP(a); lbStatusIP != nil && lbIP != nil && vip.IsIPv4(a) == isLbIPv4 && !lbIP.Equal(lbStatusIP) {
-				return []string{s.Spec.LoadBalancerIP}
-			}
-		}
-		return lbStatusAddresses
+	if len(lbStatusAddresses) > 0 || len(lbStatusHostnames) > 0 {
+		return lbStatusAddresses, lbStatusHostnames
 	}
 
 	if s.Spec.LoadBalancerIP != "" {
-		return []string{s.Spec.LoadBalancerIP}
+		return []string{s.Spec.LoadBalancerIP}, []string{}
 	}
 
-	return []string{}
+	return []string{}, []string{}
 }
 
 func FindServiceInstance(svc *v1.Service, instances []*Instance) *Instance {
-	log.Debug("finding service", "UID", svc.UID)
+	log.Debug("finding service", "namespace", svc.Namespace, "name", svc.Name, "UID", svc.UID)
 	for i := range instances {
-		log.Debug("saved service", "instance", i, "UID", instances[i].ServiceSnapshot.UID)
 		if instances[i].ServiceSnapshot.UID == svc.UID {
 			return instances[i]
 		}
 	}
+	log.Debug("instance not found", "namespace", svc.Namespace, "name", svc.Name, "UID", svc.UID)
 	return nil
 }

pkg/kubevip/annotations.go

@@ -31,19 +31,64 @@ const (
 	// Networks that we wont Egress for
 	EgressDeniedNetworks = "kube-vip.io/egress-denied-networks"
 
+	// EgressNoInternalTraffic, when enabled will prevent any internal traffic from being SNATed to the egress VIP, even if the internal SNAT rule is enabled
+	EgressNoInternalTraffic = "kube-vip.io/egress-no-internal-traffic"
+
+	// EgressDetectAPIServer, this will attempt to detect the API server and add it to the allowed networks for egress
+	EgressDetectAPIServer = "kube-vip.io/egress-detect-api-server"
+
 	// The current active endpoint(pod) for the Egress VIP
 	ActiveEndpoint = "kube-vip.io/active-endpoint"
 
 	// The current active endpoint(pod) for the Egress VIP (v6)
 	ActiveEndpointIPv6 = "kube-vip.io/active-endpoint-ipv6"
 
+	// The nftables egress table base name that owns this Service's SNAT chain
+	EgressNftablesTable = "kube-vip.io/egress-nftables-table"
+
 	// Flush the conntrack rules (remove existing sessions) once Egress is configured
 	FlushContrack = "kube-vip.io/flush-conntrack"
 
+	// Configure LoadBalancer IPs instead of relying on a controller
 	LoadbalancerIPAnnotation = "kube-vip.io/loadbalancerIPs"
-	LoadbalancerHostname     = "kube-vip.io/loadbalancerHostname"
-	ServiceInterface         = "kube-vip.io/serviceInterface"
-	UpnpEnabled              = "kube-vip.io/forwardUPNP"
+
+	// Ignore the LoadBalancer Service
+	LoadbalancerIgnore = "kube-vip.io/ignore"
+
+	// Used to configure DHCP with a Hostname
+	LoadbalancerHostname = "kube-vip.io/loadbalancerHostname"
+
+	// Define an interface name to bind the address of the LoadBalancer to
+	ServiceInterface = "kube-vip.io/serviceInterface"
+
+	// Specify VLAN subinterface for service (e.g. eth0.200)
+	ServiceVlan = "kube-vip.io/serviceVLAN"
+
+	ServiceSecurityIgnore = "kube-vip.io/ignore-service-security"
+
+	// Enable UPNP on a Service
+	UpnpEnabled = "kube-vip.io/forwardUPNP"
+
+	// Set the UPNP lease duration for a specific service using duration format (e.g., "30s", "1h")
+	UpnpLeaseDuration = "kube-vip.io/upnp-lease-duration"
 
 	RPFilter = "kube-vip.io/rp_filter" // Set the return path filter for a specific service interface
+
+	// Name of the service lease object
+	ServiceLease = "kube-vip.io/leaseName"
+
+	// Forces kube-vip to use per service election for this particular service
+	ForcePerServiceElection = "kube-vip.io/forcePerServiceElection"
+
+	// Allow service reconciliation even when no endpoints are present (Cluster policy only)
+	AllowReconcileWithoutEndpoints = "kube-vip.io/allow-reconcile-without-endpoints"
+
+	// Enable DDNS for the service
+	ServiceDDNS = "kube-vip.io/ddns"
+
+	// Forces kube-vip to use the specified veth interface when DHCP is being used for a service
+	MacvlanName = "kube-vip.io/macvlanName"
+
+	// Set the BROADCAST flag in DHCP DISCOVER/REQUEST packets
+	DHCPBroadcast = "kube-vip.io/dhcp-broadcast"
 )

pkg/kubevip/config_bgp.go

@@ -0,0 +1,335 @@
+package kubevip
+
+import (
+	"fmt"
+	"net"
+	"strconv"
+	"strings"
+
+	"github.com/kube-vip/kube-vip/pkg/utils"
+	api "github.com/osrg/gobgp/v4/api"
+
+	"github.com/vishvananda/netlink"
+)
+
+// Peer defines a BGP Peer
+type BGPPeer struct {
+	Address      string
+	Port         uint16
+	Interface    string
+	AS           uint32
+	Password     string
+	MultiHop     bool
+	MpbgpNexthop string
+	MpbgpIPv4    string
+	MpbgpIPv6    string
+
+	// BFD Configuration
+	BFDEnabled          bool
+	BFDReceiveInterval  uint32
+	BFDTransmitInterval uint32
+	BFDDetectMultiplier uint32
+}
+
+// Config defines the BGP server configuration
+type BGPConfig struct {
+	AS           uint32
+	RouterID     string
+	SourceIP     string
+	SourceIF     string
+	MpbgpNexthop string
+	MpbgpIPv4    string
+	MpbgpIPv6    string
+
+	HoldTime          uint64
+	KeepaliveInterval uint64
+
+	Peers []BGPPeer
+
+	Zebra ZebraConfig
+}
+
+// Defines Zebra connection configuration. More on the topic - https://github.com/osrg/gobgp/blob/master/docs/sources/zebra.md#configuration
+type ZebraConfig struct {
+	Enabled      bool
+	URL          string
+	Version      uint32
+	SoftwareName string
+}
+
+// BGP Peer layout is as follows:
+// <address>:<AS>:<password>:<multihop>:<port>:<optional mpbgp options>:<BFD options>
+
+// <address> - IP address of the peer. For IPv6 addresses, the address should be enclosed in square brackets (e.g. [fd00:100:64::2]). For unnumbered peers, the address should be prefixed with "unnumbered:" followed by the interface name (e.g. unnumbered:eth0).
+// <AS> - Autonomous System number of the peer (e.g. 65000)
+// <password> - Optional password for BGP authentication (e.g. secret)
+// <multihop> - Optional flag to indicate if this is a multihop peer (true/false, default: false)
+// <port> - Optional BGP port number (default: 179)
+// <optional mpbgp options> - Optional MP-BGP parameters in the format of key=value pairs separated by ';' (e.g. mpbgp_nexthop=auto_sourceif;mpbgp_ipv4=)
+// <BFD options> - Optional BFD parameters (if any) in the format of semicolon-separated values (enable, receive_interval, transmit_interval, detect_multiplier) (e.g. true;300;300;3)
+
+// ParseBGPPeerConfig - take a string and parses it into an array of peers
+func ParseBGPPeerConfig(config string) (bgpPeers []BGPPeer, err error) {
+	peers := strings.Split(config, ",")
+	if len(peers) == 0 || config == "" {
+		return nil, fmt.Errorf("no BGP Peer configurations found")
+	}
+
+	for x := range peers {
+		peerStr := peers[x]
+		if peerStr == "" {
+			continue
+		}
+
+		// Look at address peer
+		isV6Peer := peerStr[0] == '['
+		isUnnumberedPeer := strings.HasPrefix(peerStr, "unnumbered:")
+
+		address := ""
+		if isV6Peer {
+			addressEndPos := strings.IndexByte(peerStr, ']')
+			if addressEndPos == -1 {
+				return nil, fmt.Errorf("no matching ] found for IPv6 BGP Peer")
+			}
+			address = peerStr[1:addressEndPos]
+			peerStr = peerStr[addressEndPos+1:]
+		} else if isUnnumberedPeer {
+			unnumberedEndPos := strings.IndexByte(peerStr, ':')
+			peerStr = peerStr[unnumberedEndPos+1:]
+		}
+
+		peer := strings.Split(peerStr, ":")
+		if len(peer) < 2 && !isUnnumberedPeer {
+			return nil, fmt.Errorf("mandatory peering params <host>:<AS> incomplete")
+		}
+
+		iface := ""
+		if isUnnumberedPeer {
+			iface = peer[0]
+		} else if !isV6Peer {
+			address = peer[0]
+		}
+
+		// Look at peer[1] for AS number
+		var ASNumber uint64
+		if len(peer) >= 2 {
+			ASNumber, err = strconv.ParseUint(peer[1], 10, 32)
+			if err != nil {
+				return nil, fmt.Errorf("BGP Peer AS format error [%s]", peer[1])
+			}
+		}
+
+		// Look at peer[2] for password
+		password := ""
+		if len(peer) >= 3 {
+			password = peer[2]
+		}
+
+		// Look at peer[3] for multihop
+		multiHop := false
+		if len(peer) >= 4 && peer[3] != "" {
+			multiHop, err = strconv.ParseBool(peer[3])
+			if err != nil {
+				return nil, fmt.Errorf("BGP MultiHop format error (true/false) [%s]", peer[3])
+			}
+		}
+
+		// Look at peer[4] for BGP port
+		var port uint64
+		if len(peer) >= 5 {
+			if peer[4] == "" {
+				port = 179
+			} else {
+				port, err = strconv.ParseUint(peer[4], 10, 16)
+				if err != nil {
+					return nil, fmt.Errorf("BGP Peer Port format error [%s]", peer[4])
+				}
+			}
+		} else if !isUnnumberedPeer {
+			port = 179
+		}
+
+		// Look at peer[5] for optional MP-BGP parameters
+		var mpbgpNexthop, mpbgpIPv4, mpbgpIPv6 string
+
+		if len(peer) >= 6 && peer[5] != "" {
+			configData := strings.Split(peer[5], ";")
+			for _, cfg := range configData {
+				c := strings.Split(cfg, "=")
+				if len(c) < 2 {
+					return nil, fmt.Errorf("peer configuration parameter '%s' is missing a value (expected key=value)", c[0])
+				}
+				switch c[0] {
+				case "mpbgp_nexthop":
+					mpbgpNexthop = c[1]
+				case "mpbgp_ipv4":
+					mpbgpIPv4 = c[1]
+				case "mpbgp_ipv6":
+					mpbgpIPv6 = c[1]
+				default:
+					return nil, fmt.Errorf("peer configuration parameter '%s' is not supported", c[0])
+				}
+			}
+
+		}
+
+		// Look at peer[6] for optional BFD parameters (if any)
+		bfdEnabled := false
+		bfdReceiveInterval := uint64(300)
+		bfdTransmitInterval := uint64(300)
+		bfdDetectMultiplier := uint64(3)
+
+		if len(peer) >= 7 && peer[6] != "" {
+			c := strings.Split(peer[6], ";")
+			if len(c) < 4 {
+				return nil, fmt.Errorf("BFD configuration error: at least 4 parameters are required (enable, receive_interval, transmit_interval, detect_multiplier) [%s]", peer[6])
+			}
+			bfdEnabled, err = strconv.ParseBool(c[0])
+			if err != nil {
+				return nil, fmt.Errorf("BFD configuration error: invalid value for bfd_enabled (true/false) [%s]", c[0])
+			}
+
+			if c[1] != "" {
+				bfdReceiveInterval, err = strconv.ParseUint(c[1], 10, 32)
+				if err != nil {
+					return nil, fmt.Errorf("BFD configuration error: invalid value for bfd_receive_interval [%s]", c[1])
+				}
+			}
+
+			if c[2] != "" {
+				bfdTransmitInterval, err = strconv.ParseUint(c[2], 10, 32)
+				if err != nil {
+					return nil, fmt.Errorf("BFD configuration error: invalid value for bfd_transmit_interval [%s]", c[2])
+				}
+			}
+			if c[3] != "" {
+				bfdDetectMultiplier, err = strconv.ParseUint(c[3], 10, 32)
+				if err != nil {
+					return nil, fmt.Errorf("BFD configuration error: invalid value for bfd_detect_multiplier [%s]", c[3])
+				}
+			}
+		}
+
+		peerConfig := BGPPeer{
+			Address: address,
+			//nolint:gosec // previously parsed into uint32
+			AS:                  uint32(ASNumber),
+			Port:                uint16(port),
+			Interface:           iface,
+			Password:            password,
+			MultiHop:            multiHop,
+			MpbgpNexthop:        mpbgpNexthop,
+			MpbgpIPv4:           mpbgpIPv4,
+			MpbgpIPv6:           mpbgpIPv6,
+			BFDEnabled:          bfdEnabled,
+			BFDReceiveInterval:  uint32(bfdReceiveInterval),
+			BFDTransmitInterval: uint32(bfdTransmitInterval),
+			BFDDetectMultiplier: uint32(bfdDetectMultiplier),
+		}
+
+		bgpPeers = append(bgpPeers, peerConfig)
+	}
+	return
+}
+
+func (p *BGPPeer) FindMpbgpAddresses(ap *api.Peer, server *BGPConfig) (string, string, error) {
+	var ipv4Address, ipv6Address string
+
+	mode := server.MpbgpNexthop
+	if p.MpbgpNexthop != "" {
+		mode = p.MpbgpNexthop
+	}
+
+	switch mode {
+	case "fixed":
+		ap.Transport.LocalAddress = server.SourceIP
+
+		ipv4 := server.MpbgpIPv4
+		if p.MpbgpIPv4 != "" {
+			ipv4 = p.MpbgpIPv4
+		}
+
+		ipv6 := server.MpbgpIPv6
+		if p.MpbgpIPv6 != "" {
+			ipv6 = p.MpbgpIPv6
+		}
+
+		if ipv4 == "" && ipv6 == "" {
+			return "", "", fmt.Errorf("to use MP-BGP with fixed address at least one IPv4 or IPv6 address has to be provided [current - IPv4: %s, IPv6: %s]",
+				ipv4, ipv6)
+		}
+
+		if ipv4 != "" {
+			if net.ParseIP(ipv4) == nil {
+				return "", "", fmt.Errorf("provided address '%s' is not a valid IPv4 address", ipv4)
+			}
+		}
+		if ipv6 != "" {
+			if net.ParseIP(ipv6) == nil {
+				return "", "", fmt.Errorf("provided address '%s' is not a valid IPv6 address", ipv6)
+			}
+		}
+
+		ipv4Address = ipv4
+		ipv6Address = ipv6
+	case "auto_sourceip":
+		ap.Transport.LocalAddress = server.SourceIP
+
+		// Resolve the local interface by SourceIP
+		iface, err := utils.GetInterfaceByIP(server.SourceIP)
+		if err != nil {
+			return "", "", fmt.Errorf("failed to get interface by IP: %v", err)
+		}
+
+		if utils.IsIPv4(server.SourceIP) {
+			// Get the non link-local IPv6 address on that interface
+			ipv6Address, err = utils.GetNonLinkLocalIP(iface, netlink.FAMILY_V6)
+			if err != nil {
+				return "", "", fmt.Errorf("failed to get non link-local IPv6 address: %v", err)
+			}
+		} else {
+			// Get the non link-local IPv4 address on that interface
+			ipv4Address, err = utils.GetNonLinkLocalIP(iface, netlink.FAMILY_V4)
+			if err != nil {
+				return "", "", fmt.Errorf("failed to get non link-local IPv4 address: %v", err)
+			}
+		}
+	case "auto_sourceif":
+		ap.Transport.BindInterface = server.SourceIF
+
+		iface, err := netlink.LinkByName(server.SourceIF)
+		if err != nil {
+			return "", "", fmt.Errorf("failed to get interface by name: %v", err)
+		}
+
+		// Get the non link-local IPv4 address on that interface
+		ipv4Address, err = utils.GetNonLinkLocalIP(&iface, netlink.FAMILY_V4)
+		if err != nil {
+			return "", "", fmt.Errorf("failed to get non link-local IPv4 address: %v", err)
+		}
+
+		// Get the non link-local IPv6 address on that interface
+		ipv6Address, err = utils.GetNonLinkLocalIP(&iface, netlink.FAMILY_V6)
+		if err != nil {
+			return "", "", fmt.Errorf("failed to get non link-local IPv6 address: %v", err)
+		}
+	default:
+		return "", "", fmt.Errorf("option %q for MP-BPG nexthop is not supported", mode)
+	}
+
+	return ipv4Address, ipv6Address, nil
+}
+
+func (p *BGPPeer) SetMpbgpOptions(server *BGPConfig) {
+	if p.MpbgpNexthop == "" {
+		p.MpbgpNexthop = server.MpbgpNexthop
+	}
+
+	if p.MpbgpIPv4 == "" {
+		p.MpbgpIPv4 = server.MpbgpIPv4
+	}
+
+	if p.MpbgpIPv6 == "" {
+		p.MpbgpIPv6 = server.MpbgpIPv6
+	}
+}

pkg/kubevip/config_bgp_test.go

@@ -0,0 +1,116 @@
+package kubevip
+
+import (
+	"reflect"
+	"testing"
+)
+
+func TestParseBGPPeerConfig(t *testing.T) {
+	type args struct {
+		config string
+	}
+	tests := []struct {
+		name         string
+		args         args
+		wantBgpPeers []BGPPeer
+		wantErr      bool
+	}{
+
+		{
+			name: "IPv4, default port",
+			args: args{config: "192.168.0.10:65000::false,192.168.0.11:65000::false"},
+			wantBgpPeers: []BGPPeer{
+				{Address: "192.168.0.10", Port: 179, AS: 65000, MultiHop: false, BFDEnabled: false, BFDReceiveInterval: 300, BFDTransmitInterval: 300, BFDDetectMultiplier: 3},
+				{Address: "192.168.0.11", Port: 179, AS: 65000, MultiHop: false, BFDEnabled: false, BFDReceiveInterval: 300, BFDTransmitInterval: 300, BFDDetectMultiplier: 3},
+			},
+		},
+		{
+			name: "IPv4, different port",
+			args: args{config: "192.168.0.10:65000::false:180,192.168.0.11:65000::false:190"},
+			wantBgpPeers: []BGPPeer{
+				{Address: "192.168.0.10", Port: 180, AS: 65000, MultiHop: false, BFDEnabled: false, BFDReceiveInterval: 300, BFDTransmitInterval: 300, BFDDetectMultiplier: 3},
+				{Address: "192.168.0.11", Port: 190, AS: 65000, MultiHop: false, BFDEnabled: false, BFDReceiveInterval: 300, BFDTransmitInterval: 300, BFDDetectMultiplier: 3},
+			},
+		},
+		{
+			name: "IPv6, multi-protocol",
+			args: args{config: "[fd00:1111:2222:3333:c7d9:7235:6bf7:5d52]:65501::false::mpbgp_nexthop=auto_sourceif"},
+			wantBgpPeers: []BGPPeer{
+				{Address: "fd00:1111:2222:3333:c7d9:7235:6bf7:5d52", Port: 179, AS: 65501, MultiHop: false, MpbgpNexthop: "auto_sourceif", BFDEnabled: false, BFDReceiveInterval: 300, BFDTransmitInterval: 300, BFDDetectMultiplier: 3},
+			},
+		},
+		{
+			name: "IPv6, multi-protocol, BFD, no multi-protocol options",
+			args: args{config: "[fd00:1111:2222:3333:c7d9:7235:6bf7:5d52]:65501::false:::true;300;300;3"},
+			wantBgpPeers: []BGPPeer{
+				{Address: "fd00:1111:2222:3333:c7d9:7235:6bf7:5d52", Port: 179, AS: 65501, MultiHop: false, BFDEnabled: true, BFDReceiveInterval: 300, BFDTransmitInterval: 300, BFDDetectMultiplier: 3},
+			},
+		},
+		{
+			name: "IPv6, multi-protocol, BFD",
+			args: args{config: "[fd00:1111:2222:3333:c7d9:7235:6bf7:5d52]:65501::false::mpbgp_nexthop=auto_sourceif:true;300;300;3"},
+			wantBgpPeers: []BGPPeer{
+				{Address: "fd00:1111:2222:3333:c7d9:7235:6bf7:5d52", Port: 179, AS: 65501, MultiHop: false, MpbgpNexthop: "auto_sourceif", BFDEnabled: true, BFDReceiveInterval: 300, BFDTransmitInterval: 300, BFDDetectMultiplier: 3},
+			},
+		},
+		{
+			name: "IPv6 bracketed, with password and multihop",
+			args: args{config: "[fd00:100:64::2]:65000:secret:true"},
+			wantBgpPeers: []BGPPeer{
+				{Address: "fd00:100:64::2", Port: 179, AS: 65000, Password: "secret", MultiHop: true, BFDEnabled: false, BFDReceiveInterval: 300, BFDTransmitInterval: 300, BFDDetectMultiplier: 3},
+			},
+		},
+		{
+			name: "IPv6 bracketed, empty fields",
+			args: args{config: "[fd00:100:64::2]:65000::false"},
+			wantBgpPeers: []BGPPeer{
+				{Address: "fd00:100:64::2", Port: 179, AS: 65000, MultiHop: false, BFDEnabled: false, BFDReceiveInterval: 300, BFDTransmitInterval: 300, BFDDetectMultiplier: 3},
+			},
+		},
+		{
+			name: "Unnumbered",
+			args: args{config: "unnumbered:eth0,unnumbered:eth1:65000::true::mpbgp_nexthop=auto_sourceif"},
+			wantBgpPeers: []BGPPeer{
+				{Interface: "eth0", MultiHop: false, BFDEnabled: false, BFDReceiveInterval: 300, BFDTransmitInterval: 300, BFDDetectMultiplier: 3},
+				{Interface: "eth1", Port: 179, AS: 65000, MultiHop: true, MpbgpNexthop: "auto_sourceif", BFDEnabled: false, BFDReceiveInterval: 300, BFDTransmitInterval: 300, BFDDetectMultiplier: 3},
+			},
+		},
+		{
+			name:    "Completely empty config",
+			args:    args{config: ""},
+			wantErr: true,
+		},
+		{
+			name:    "Completely empty config (but with the seperators)",
+			args:    args{config: ":::::::"},
+			wantErr: true,
+		},
+		{
+			name:    "Malformed parameter (no value)",
+			args:    args{config: "1.2.3.4:65000/mpbgp_nexthop"},
+			wantErr: true,
+		},
+		{
+			name:    "Unsupported parameter",
+			args:    args{config: "1.2.3.4:65000;unknown=value"},
+			wantErr: true,
+		},
+		{
+			name:    "Malformed IPv6 (no matching bracket)",
+			args:    args{config: "[fd00:100:64::2:65000"},
+			wantErr: true,
+		},
+	}
+	for _, tt := range tests {
+		t.Run(tt.name, func(t *testing.T) {
+			gotBgpPeers, err := ParseBGPPeerConfig(tt.args.config)
+			if (err != nil) != tt.wantErr {
+				t.Errorf("ParseBGPPeerConfig() error = \n%v, wantErr \n%v %v", err, tt.wantErr, gotBgpPeers)
+				return
+			}
+			if !reflect.DeepEqual(gotBgpPeers, tt.wantBgpPeers) {
+				t.Errorf("ParseBGPPeerConfig() = \n%v, want \n%v", gotBgpPeers, tt.wantBgpPeers)
+			}
+		})
+	}
+}

pkg/kubevip/config_environment.go

@@ -6,11 +6,14 @@ import (
 	"math"
 	"math/bits"
 	"os"
+	"path/filepath"
 	"strconv"
 	"strings"
 
-	"github.com/kube-vip/kube-vip/pkg/bgp"
+	"github.com/kube-vip/kube-vip/pkg/debouncer"
 	"github.com/kube-vip/kube-vip/pkg/detector"
+	"github.com/kube-vip/kube-vip/pkg/utils"
+	"sigs.k8s.io/yaml"
 )
 
 // ParseEnvironment - will popultate the configuration from environment variables
@@ -24,11 +27,18 @@ func ParseEnvironment(c *Config) error {
 	if env != "" {
 		logLevel, err := strconv.ParseInt(env, 10, 32)
 		if err != nil {
-			panic("Unable to parse environment variable [vip_loglevel], should be int")
+			return fmt.Errorf("unable to parse environment variable [vip_loglevel], should be int: %w", err)
 		}
 		c.Logging = int32(logLevel)
 	}
 
+	if env = os.Getenv(instanceName); env == "" {
+		env = os.Getenv(strings.ToUpper(instanceName))
+	}
+	if env != "" {
+		c.InstanceName = env
+	}
+
 	// Find interface
 	env = os.Getenv(vipInterface)
 	if env != "" {
@@ -44,12 +54,39 @@ func ParseEnvironment(c *Config) error {
 		c.LoInterfaceGlobalScope = b
 	}
 
+	env = os.Getenv(vipLoseLeadership)
+	if env != "" {
+		b, err := strconv.ParseBool(env)
+		if err != nil {
+			return err
+		}
+		c.LoseLeadership = b
+	}
+
+	env = os.Getenv(vipLoseLeadershipTimeoutSeconds)
+	if env != "" {
+		i, err := strconv.ParseInt(env, 10, 32)
+		if err != nil {
+			return fmt.Errorf("parsing env var %s (value: %s): %w", vipLoseLeadershipTimeoutSeconds, env, err)
+		}
+		c.LoseLeadershipTimeoutSeconds = int(i)
+	}
 	// Find (services) interface
 	env = os.Getenv(vipServicesInterface)
 	if env != "" {
 		c.ServicesInterface = env
 	}
 
+	// Tolerate a down interface
+	env = os.Getenv(vipAllowInterfaceNotUp)
+	if env != "" {
+		b, err := strconv.ParseBool(env)
+		if err != nil {
+			return err
+		}
+		c.AllowInterfaceNotUp = b
+	}
+
 	// Find Kubernetes Leader Election configuration
 	env = os.Getenv(vipLeaderElection)
 	if env != "" {
@@ -290,6 +327,18 @@ func ParseEnvironment(c *Config) error {
 		c.ArpBroadcastRate = 3000
 	}
 
+	// Determine if VIP should be preserved on leadership loss
+	// true: VIP addresses remain on interface, only ARP/NDP broadcasting stops
+	// false (default): VIP addresses are deleted on leadership loss (legacy behavior)
+	env = os.Getenv(vipPreserveOnLeadershipLoss)
+	if env != "" {
+		b, err := strconv.ParseBool(env)
+		if err != nil {
+			return err
+		}
+		c.PreserveVIPOnLeadershipLoss = b
+	}
+
 	// Wireguard Mode
 	env = os.Getenv(vipWireguard)
 	if env != "" {
@@ -357,12 +406,46 @@ func ParseEnvironment(c *Config) error {
 		c.CleanRoutingTable = b
 	}
 
+	// Skip Duplicate Address Detection when adding the VIP address
+	env = os.Getenv(vipSkipDAD)
+	if env != "" {
+		b, err := strconv.ParseBool(env)
+		if err != nil {
+			return err
+		}
+		c.SkipDAD = b
+	}
+
 	// DNS mode
 	env = os.Getenv(dnsMode)
 	if env != "" {
 		c.DNSMode = env
 	}
 
+	// DHCP mode
+	env = os.Getenv(dhcpMode)
+	if env != "" {
+		c.DHCPMode = env
+	} else {
+		if c.DNSMode != "first" {
+			c.DHCPMode = c.DNSMode
+		} else {
+			c.DHCPMode = strings.ToLower(utils.IPv4Family)
+		}
+	}
+
+	// DHCP backoff attempts
+	env = os.Getenv(dhcpBackoffAttempts)
+	if env != "" {
+		tmp, err := strconv.ParseInt(env, 10, 32)
+		if err != nil {
+			return err
+		}
+		if tmp >= 0 {
+			c.DHCPBackoffAttempts = uint(tmp)
+		}
+	}
+
 	// Disable updates for services (status.LoadBalancer.Ingress will not be updated)
 	env = os.Getenv(disableServiceUpdates)
 	if env != "" {
@@ -383,6 +466,15 @@ func ParseEnvironment(c *Config) error {
 		c.EnableBGP = b
 	}
 
+	env = os.Getenv(bgpAttachIPToInterface)
+	if env != "" {
+		b, err := strconv.ParseBool(env)
+		if err != nil {
+			return err
+		}
+		c.BGPAttachIPToInterface = b
+	}
+
 	// BGP Router interface determines an interface that we can use to find an address for
 	env = os.Getenv(bgpRouterInterface)
 	if env != "" {
@@ -422,7 +514,7 @@ func ParseEnvironment(c *Config) error {
 	// Peer AS
 	env = os.Getenv(bgpPeers)
 	if env != "" {
-		peers, err := bgp.ParseBGPPeerConfig(env)
+		peers, err := ParseBGPPeerConfig(env)
 		if err != nil {
 			return err
 		}
@@ -501,6 +593,40 @@ func ParseEnvironment(c *Config) error {
 		c.BGPConfig.KeepaliveInterval = u64
 	}
 
+	// BGP health check options
+	env = os.Getenv(controlPlaneHealthCheckAddress)
+	if env != "" {
+		c.ControlPlaneHealthCheck.Address = env
+	}
+	env = os.Getenv(controlPlaneHealthCheckPeriodSeconds)
+	if env != "" {
+		i, err := strconv.ParseInt(env, 10, 32)
+		if err != nil {
+			return fmt.Errorf("parsing env var %s (value: %s): %w", controlPlaneHealthCheckPeriodSeconds, env, err)
+		}
+		c.ControlPlaneHealthCheck.PeriodSeconds = int(i)
+	}
+	env = os.Getenv(controlPlaneHealthCheckTimeoutSeconds)
+	if env != "" {
+		i, err := strconv.ParseInt(env, 10, 32)
+		if err != nil {
+			return fmt.Errorf("parsing env var %s (value: %s): %w", controlPlaneHealthCheckTimeoutSeconds, env, err)
+		}
+		c.ControlPlaneHealthCheck.TimeoutSeconds = int(i)
+	}
+	env = os.Getenv(controlPlaneHealthCheckFailureThreshold)
+	if env != "" {
+		i, err := strconv.ParseInt(env, 10, 32)
+		if err != nil {
+			return fmt.Errorf("parsing env var %s (value: %s): %w", controlPlaneHealthCheckFailureThreshold, env, err)
+		}
+		c.ControlPlaneHealthCheck.FailureThreshold = int(i)
+	}
+	env = os.Getenv(controlPlaneHealthCheckCAPath)
+	if env != "" {
+		c.ControlPlaneHealthCheck.CAPath = env
+	}
+
 	env = os.Getenv(zebraEnable)
 	if env != "" {
 		result, err := strconv.ParseBool(env)
@@ -601,6 +727,25 @@ func ParseEnvironment(c *Config) error {
 		c.EgressWithNftables = b
 	}
 
+	env = os.Getenv(perServiceElectionOnDemand)
+	if env != "" {
+		b, err := strconv.ParseBool(env)
+		if err != nil {
+			return err
+		}
+		c.PerServiceElectionOnDemand = b
+	}
+
+	// if this is set then we're enabling the internal SNAT rule that kube-vip adds to the egress chain
+	env = os.Getenv(egressEnableInternalSNAT)
+	if env != "" {
+		b, err := strconv.ParseBool(env)
+		if err != nil {
+			return err
+		}
+		c.EnableInternalSNAT = b
+	}
+
 	// check to see if we're using a specific path to the Kubernetes config file
 	env = os.Getenv(k8sConfigFile)
 	if env != "" {
@@ -667,5 +812,332 @@ func ParseEnvironment(c *Config) error {
 		c.EgressClean = b
 	}
 
+	// check for configuration file path
+	env = os.Getenv(configFile)
+	if env != "" {
+		c.ConfigFile = env
+	}
+
 	return nil
 }
+
+// LoadConfigFromFile loads configuration from a JSON or YAML file
+func LoadConfigFromFile(configFilePath string) (*Config, error) {
+	if configFilePath == "" {
+		return nil, fmt.Errorf("config file path is empty")
+	}
+
+	// Check if file exists
+	if _, err := os.Stat(configFilePath); os.IsNotExist(err) {
+		return nil, fmt.Errorf("config file does not exist: %s", configFilePath)
+	}
+
+	// Read file content
+	data, err := os.ReadFile(configFilePath)
+	if err != nil {
+		return nil, fmt.Errorf("failed to read config file %s: %v", configFilePath, err)
+	}
+
+	var config Config
+	ext := strings.ToLower(filepath.Ext(configFilePath))
+
+	switch ext {
+	case ".json":
+		err = json.Unmarshal(data, &config)
+		if err != nil {
+			return nil, fmt.Errorf("failed to parse JSON config file %s: %v", configFilePath, err)
+		}
+	case ".yaml", ".yml":
+		err = yaml.Unmarshal(data, &config)
+		if err != nil {
+			return nil, fmt.Errorf("failed to parse YAML config file %s: %v", configFilePath, err)
+		}
+	default:
+		return nil, fmt.Errorf("unsupported config file format %s. Supported formats: .json, .yaml, .yml", ext)
+	}
+
+	return &config, nil
+}
+
+// MergeConfigFromFile merges configuration loaded from file with existing config
+// Priority: command line flags > environment variables > config file
+func MergeConfigFromFile(c *Config, configFilePath string) error {
+	if configFilePath == "" {
+		return nil // No config file specified, nothing to merge
+	}
+
+	fileConfig, err := LoadConfigFromFile(configFilePath)
+	if err != nil {
+		return err
+	}
+
+	// Merge file config with existing config
+	// Only set values from file if they haven't been set by flags or env vars
+	mergeConfigValues(c, fileConfig)
+
+	return nil
+}
+
+// mergeConfigValues merges values from fileConfig into baseConfig
+// Only overwrites zero values in baseConfig
+func mergeConfigValues(baseConfig, fileConfig *Config) {
+	// Basic configuration
+	if baseConfig.Logging == 0 && fileConfig.Logging != 0 {
+		baseConfig.Logging = fileConfig.Logging
+	}
+
+	// Network configuration
+	if baseConfig.Interface == "" && fileConfig.Interface != "" {
+		baseConfig.Interface = fileConfig.Interface
+	}
+	if baseConfig.ServicesInterface == "" && fileConfig.ServicesInterface != "" {
+		baseConfig.ServicesInterface = fileConfig.ServicesInterface
+	}
+	if baseConfig.VIP == "" && fileConfig.VIP != "" {
+		baseConfig.VIP = fileConfig.VIP
+	}
+	if baseConfig.VIPSubnet == "" && fileConfig.VIPSubnet != "" {
+		baseConfig.VIPSubnet = fileConfig.VIPSubnet
+	}
+	if baseConfig.Address == "" && fileConfig.Address != "" {
+		baseConfig.Address = fileConfig.Address
+	}
+	if baseConfig.Port == 0 && fileConfig.Port != 0 {
+		baseConfig.Port = fileConfig.Port
+	}
+	if baseConfig.NodeName == "" && fileConfig.NodeName != "" {
+		baseConfig.NodeName = fileConfig.NodeName
+	}
+
+	// Boolean flags - only merge if not explicitly set
+	if !baseConfig.EnableARP && fileConfig.EnableARP {
+		baseConfig.EnableARP = fileConfig.EnableARP
+	}
+	if !baseConfig.EnableBGP && fileConfig.EnableBGP {
+		baseConfig.EnableBGP = fileConfig.EnableBGP
+	}
+	if !baseConfig.BGPAttachIPToInterface && fileConfig.BGPAttachIPToInterface {
+		baseConfig.BGPAttachIPToInterface = fileConfig.BGPAttachIPToInterface
+	}
+	if !baseConfig.EnableWireguard && fileConfig.EnableWireguard {
+		baseConfig.EnableWireguard = fileConfig.EnableWireguard
+	}
+	if !baseConfig.EnableRoutingTable && fileConfig.EnableRoutingTable {
+		baseConfig.EnableRoutingTable = fileConfig.EnableRoutingTable
+	}
+	if !baseConfig.EnableControlPlane && fileConfig.EnableControlPlane {
+		baseConfig.EnableControlPlane = fileConfig.EnableControlPlane
+	}
+	if !baseConfig.DetectControlPlane && fileConfig.DetectControlPlane {
+		baseConfig.DetectControlPlane = fileConfig.DetectControlPlane
+	}
+	if !baseConfig.EnableServices && fileConfig.EnableServices {
+		baseConfig.EnableServices = fileConfig.EnableServices
+	}
+	if !baseConfig.EnableServicesElection && fileConfig.EnableServicesElection {
+		baseConfig.EnableServicesElection = fileConfig.EnableServicesElection
+	}
+	if !baseConfig.EnableNodeLabeling && fileConfig.EnableNodeLabeling {
+		baseConfig.EnableNodeLabeling = fileConfig.EnableNodeLabeling
+	}
+	if !baseConfig.EnableLoadBalancer && fileConfig.EnableLoadBalancer {
+		baseConfig.EnableLoadBalancer = fileConfig.EnableLoadBalancer
+	}
+	if !baseConfig.DDNS && fileConfig.DDNS {
+		baseConfig.DDNS = fileConfig.DDNS
+	}
+	if !baseConfig.SingleNode && fileConfig.SingleNode {
+		baseConfig.SingleNode = fileConfig.SingleNode
+	}
+	if !baseConfig.StartAsLeader && fileConfig.StartAsLeader {
+		baseConfig.StartAsLeader = fileConfig.StartAsLeader
+	}
+	if !baseConfig.PreserveVIPOnLeadershipLoss && fileConfig.PreserveVIPOnLeadershipLoss {
+		baseConfig.PreserveVIPOnLeadershipLoss = fileConfig.PreserveVIPOnLeadershipLoss
+	}
+
+	// Service configuration
+	if baseConfig.Namespace == "" && fileConfig.Namespace != "" {
+		baseConfig.Namespace = fileConfig.Namespace
+	}
+	if baseConfig.ServiceNamespace == "" && fileConfig.ServiceNamespace != "" {
+		baseConfig.ServiceNamespace = fileConfig.ServiceNamespace
+	}
+	if baseConfig.ServicesLeaseName == "" && fileConfig.ServicesLeaseName != "" {
+		baseConfig.ServicesLeaseName = fileConfig.ServicesLeaseName
+	}
+	// LoadBalancer configuration
+	if baseConfig.LoadBalancerPort == 0 && fileConfig.LoadBalancerPort != 0 {
+		baseConfig.LoadBalancerPort = fileConfig.LoadBalancerPort
+	}
+	if baseConfig.LoadBalancerForwardingMethod == "" && fileConfig.LoadBalancerForwardingMethod != "" {
+		baseConfig.LoadBalancerForwardingMethod = fileConfig.LoadBalancerForwardingMethod
+	}
+	if baseConfig.LoadBalancerClassName == "" && fileConfig.LoadBalancerClassName != "" {
+		baseConfig.LoadBalancerClassName = fileConfig.LoadBalancerClassName
+	}
+
+	// Routing Table configuration
+	if baseConfig.RoutingTableID == 0 && fileConfig.RoutingTableID != 0 {
+		baseConfig.RoutingTableID = fileConfig.RoutingTableID
+	}
+	if baseConfig.RoutingTableType == 0 && fileConfig.RoutingTableType != 0 {
+		baseConfig.RoutingTableType = fileConfig.RoutingTableType
+	}
+	if baseConfig.RoutingProtocol == 0 && fileConfig.RoutingProtocol != 0 {
+		baseConfig.RoutingProtocol = fileConfig.RoutingProtocol
+	}
+
+	// BGP configuration
+	mergeBGPConfig(&baseConfig.BGPConfig, &fileConfig.BGPConfig)
+
+	// Kubernetes configuration
+	if baseConfig.K8sConfigFile == "" && fileConfig.K8sConfigFile != "" {
+		baseConfig.K8sConfigFile = fileConfig.K8sConfigFile
+	}
+
+	// Leader Election configuration
+	mergeLeaderElectionConfig(&baseConfig.KubernetesLeaderElection, &fileConfig.KubernetesLeaderElection)
+
+	// BGP health check configuration
+	mergeHealthCheck(&baseConfig.ControlPlaneHealthCheck, &fileConfig.ControlPlaneHealthCheck)
+
+	// Prometheus configuration
+	if baseConfig.PrometheusHTTPServer == "" && fileConfig.PrometheusHTTPServer != "" {
+		baseConfig.PrometheusHTTPServer = fileConfig.PrometheusHTTPServer
+	}
+
+	// DNS configuration
+	if baseConfig.DNSMode == "" && fileConfig.DNSMode != "" {
+		baseConfig.DNSMode = fileConfig.DNSMode
+	}
+
+	// DHCP configuration - mode
+	if baseConfig.DHCPMode == "" && fileConfig.DHCPMode != "" {
+		baseConfig.DHCPMode = fileConfig.DHCPMode
+	}
+
+	// DHCP configuration - backoff attempts
+	if baseConfig.DHCPBackoffAttempts == DefaultDHCPBackoffAttempts && fileConfig.DHCPBackoffAttempts != DefaultDHCPBackoffAttempts {
+		baseConfig.DHCPBackoffAttempts = fileConfig.DHCPBackoffAttempts
+	}
+
+	// Health check configuration (HTTP listener for kube-vip readiness)
+	if baseConfig.HealthCheckPort == 0 && fileConfig.HealthCheckPort != 0 {
+		baseConfig.HealthCheckPort = fileConfig.HealthCheckPort
+	}
+
+	// Instance configuration
+	if baseConfig.InstanceName == "" && fileConfig.InstanceName != "" {
+		baseConfig.InstanceName = fileConfig.InstanceName
+	}
+
+	// Egress configuration
+	if baseConfig.EgressPodCidr == "" && fileConfig.EgressPodCidr != "" {
+		baseConfig.EgressPodCidr = fileConfig.EgressPodCidr
+	}
+	if baseConfig.EgressServiceCidr == "" && fileConfig.EgressServiceCidr != "" {
+		baseConfig.EgressServiceCidr = fileConfig.EgressServiceCidr
+	}
+	// Mirror configuration
+	if baseConfig.MirrorDestInterface == "" && fileConfig.MirrorDestInterface != "" {
+		baseConfig.MirrorDestInterface = fileConfig.MirrorDestInterface
+	}
+
+	// Iptables configuration
+	if baseConfig.IptablesBackend == "" && fileConfig.IptablesBackend != "" {
+		baseConfig.IptablesBackend = fileConfig.IptablesBackend
+	}
+
+	// Backend health check interval
+	if baseConfig.BackendHealthCheckInterval == 0 && fileConfig.BackendHealthCheckInterval != 0 {
+		baseConfig.BackendHealthCheckInterval = fileConfig.BackendHealthCheckInterval
+	}
+
+	// ARP broadcast rate
+	if baseConfig.ArpBroadcastRate == 0 && fileConfig.ArpBroadcastRate != 0 {
+		baseConfig.ArpBroadcastRate = fileConfig.ArpBroadcastRate
+	}
+
+	// Annotations
+	if baseConfig.Annotations == "" && fileConfig.Annotations != "" {
+		baseConfig.Annotations = fileConfig.Annotations
+	}
+
+	// Load balancers slice
+	if len(baseConfig.LoadBalancers) == 0 && len(fileConfig.LoadBalancers) > 0 {
+		baseConfig.LoadBalancers = fileConfig.LoadBalancers
+	}
+
+	// Debounce time for watch events
+	if baseConfig.DebounceTime == debouncer.DefaultTime && fileConfig.DebounceTime != debouncer.DefaultTime {
+		baseConfig.DebounceTime = fileConfig.DebounceTime
+	}
+
+	if baseConfig.LoseLeadershipTimeoutSeconds == 0 && fileConfig.LoseLeadershipTimeoutSeconds != 0 {
+		baseConfig.LoseLeadershipTimeoutSeconds = fileConfig.LoseLeadershipTimeoutSeconds
+	}
+}
+
+// mergeBGPConfig merges BGP configuration
+func mergeBGPConfig(base, file *BGPConfig) {
+	if base.RouterID == "" && file.RouterID != "" {
+		base.RouterID = file.RouterID
+	}
+	if base.AS == 0 && file.AS != 0 {
+		base.AS = file.AS
+	}
+	if base.SourceIF == "" && file.SourceIF != "" {
+		base.SourceIF = file.SourceIF
+	}
+	if base.SourceIP == "" && file.SourceIP != "" {
+		base.SourceIP = file.SourceIP
+	}
+	if base.HoldTime == 0 && file.HoldTime != 0 {
+		base.HoldTime = file.HoldTime
+	}
+	if base.KeepaliveInterval == 0 && file.KeepaliveInterval != 0 {
+		base.KeepaliveInterval = file.KeepaliveInterval
+	}
+	if len(base.Peers) == 0 && len(file.Peers) > 0 {
+		base.Peers = file.Peers
+	}
+}
+
+// mergeLeaderElectionConfig merges leader election configuration
+func mergeLeaderElectionConfig(base, file *KubernetesLeaderElection) {
+	if base.LeaseName == "" && file.LeaseName != "" {
+		base.LeaseName = file.LeaseName
+	}
+	if base.LeaseDuration == 0 && file.LeaseDuration != 0 {
+		base.LeaseDuration = file.LeaseDuration
+	}
+	if base.RenewDeadline == 0 && file.RenewDeadline != 0 {
+		base.RenewDeadline = file.RenewDeadline
+	}
+	if base.RetryPeriod == 0 && file.RetryPeriod != 0 {
+		base.RetryPeriod = file.RetryPeriod
+	}
+	if len(base.LeaseAnnotations) == 0 && len(file.LeaseAnnotations) > 0 {
+		base.LeaseAnnotations = file.LeaseAnnotations
+	}
+}
+
+// mergeHealthCheck merges HTTP health check configuration for BGP route advertisement.
+func mergeHealthCheck(base, file *HealthCheck) {
+	if base.Address == "" && file.Address != "" {
+		base.Address = file.Address
+	}
+	if base.PeriodSeconds == 0 && file.PeriodSeconds != 0 {
+		base.PeriodSeconds = file.PeriodSeconds
+	}
+	if base.TimeoutSeconds == 0 && file.TimeoutSeconds != 0 {
+		base.TimeoutSeconds = file.TimeoutSeconds
+	}
+	if base.FailureThreshold == 0 && file.FailureThreshold != 0 {
+		base.FailureThreshold = file.FailureThreshold
+	}
+	if base.CAPath == "" && file.CAPath != "" {
+		base.CAPath = file.CAPath
+	}
+}

pkg/kubevip/config_environment_test.go

@@ -0,0 +1,41 @@
+package kubevip
+
+import (
+	"testing"
+)
+
+func TestParseEnvironmentSkipDAD(t *testing.T) {
+	cases := []struct {
+		name    string
+		value   string
+		want    bool
+		wantErr bool
+	}{
+		{name: "unset keeps default false", value: "", want: false},
+		{name: "true enables", value: "true", want: true},
+		{name: "false disables", value: "false", want: false},
+		{name: "garbage errors", value: "not-a-bool", wantErr: true},
+	}
+
+	for _, tc := range cases {
+		t.Run(tc.name, func(t *testing.T) {
+			if tc.value != "" {
+				t.Setenv(vipSkipDAD, tc.value)
+			}
+			c := &Config{}
+			err := ParseEnvironment(c)
+			if tc.wantErr {
+				if err == nil {
+					t.Fatal("expected an error, got nil")
+				}
+				return
+			}
+			if err != nil {
+				t.Fatalf("unexpected error: %v", err)
+			}
+			if c.SkipDAD != tc.want {
+				t.Fatalf("SkipDAD = %v, want %v", c.SkipDAD, tc.want)
+			}
+		})
+	}
+}

pkg/kubevip/config_envvar.go

@@ -9,36 +9,51 @@ const (
 	// vip_arpRate - defines the rate of gARP broadcasts
 	vipArpRate = "vip_arpRate"
 
+	// vipPreserveOnLeadershipLoss - if true, VIP addresses will remain on interface when leadership is lost
+	vipPreserveOnLeadershipLoss = "vip_preserve_on_leadership_loss"
+
 	// vipLeaderElection - defines if the kubernetes algorithm should be used
 	vipLeaderElection = "vip_leaderelection"
 
 	// vipLeaseName - defines the name of the lease lock
 	vipLeaseName = "vip_leasename"
 
-	// vipLeaderElection - defines if the kubernetes algorithm should be used
+	// vipLeaseDuration - defines how long the current leader is considered valid
 	vipLeaseDuration = "vip_leaseduration"
 
-	// vipLeaderElection - defines if the kubernetes algorithm should be used
+	// vipRenewDeadline - defines how long the leader has to renew the lease before losing leadership
 	vipRenewDeadline = "vip_renewdeadline"
 
-	// vipLeaderElection - defines if the kubernetes algorithm should be used
+	// vipRetryPeriod - defines the time between attempts to acquire/renew the lease
 	vipRetryPeriod = "vip_retryperiod"
 
-	// vipLeaderElection - defines the annotations given to the lease lock
+	// vipLeaseAnnotations - defines the annotations given to the lease lock
 	vipLeaseAnnotations = "vip_leaseannotations"
 
 	// vipLogLevel - defines the level of logging to produce (5 being the most verbose)
 	vipLogLevel = "vip_loglevel"
 
+	// instanceName identifies this kube-vip deployment when naming host-global resources
+	instanceName = "instance_name"
+
 	// vipInterface - defines the interface that the vip should bind too
 	vipInterface = "vip_interface"
 
+	// vipLoseLeadership - defines if leader should lose leadership if network interface is down
+	vipLoseLeadership = "vip_loseleadership"
+
+	// vipLoseLeadershipTimeout - defines the timeout for lose leadership
+	vipLoseLeadershipTimeoutSeconds = "vip_loseleadership_timeout_seconds"
+
 	// vipInterfaceLoGlobal - defines if the lo interface (if used) should have a global scope
 	vipInterfaceLoGlobal = "vip_interfaceloglobal"
 
 	// vipServicesInterface - defines the interface that the service vips should bind too
 	vipServicesInterface = "vip_servicesinterface"
 
+	// vipAllowInterfaceNotUp - defines if kube-vip should tolerate a down interface
+	vipAllowInterfaceNotUp = "vip_allow_interface_not_up"
+
 	// vipSubnet - defines the subnet that the vip will use
 	vipSubnet = "vip_subnet"
 
@@ -51,6 +66,11 @@ const (
 	// egressWithNftables - enables using nftables over iptables
 	egressWithNftables = "egress_withnftables"
 
+	// perServiceElectionOnDemand - enables kube-vip to use per-service election for annotated services
+	perServiceElectionOnDemand = "per_service_election_on_demand"
+
+	// egressEnableInternalSNAT - enables the internal SNAT rule that kube-vip adds to the egress chain
+	egressEnableInternalSNAT = "egress_enableinternalsnat"
 	/////////////////////////////////////
 	// TO DO:
 	// Determine how to tidy this mess up
@@ -85,6 +105,8 @@ const (
 
 	// bgpEnable defines if BGP should be enabled
 	bgpEnable = "bgp_enable"
+	// bgpAttachIPToInterface defines if BGP service VIPs should be assigned to the configured interface
+	bgpAttachIPToInterface = "bgp_attach_ip_to_interface"
 	// bgpRouterID defines the routerID for the BGP server
 	bgpRouterID = "bgp_routerid"
 	// bgpRouterInterface defines the interface that we can find the address for
@@ -109,6 +131,16 @@ const (
 	bgpHoldTime = "bgp_hold_time"
 	// bgpKeepaliveInterval defines bgp timers keepalive interval
 	bgpKeepaliveInterval = "bgp_keepalive_interval"
+	// controlPlaneHealthCheckAddress defines the URL for control-plane health checks (BGP route withdrawal)
+	controlPlaneHealthCheckAddress = "control_plane_health_check_address"
+	// controlPlaneHealthCheckPeriodSeconds defines the period between control-plane health checks
+	controlPlaneHealthCheckPeriodSeconds = "control_plane_health_check_period_seconds"
+	// controlPlaneHealthCheckTimeoutSeconds defines the timeout for each control-plane health check request
+	controlPlaneHealthCheckTimeoutSeconds = "control_plane_health_check_timeout_seconds"
+	// controlPlaneHealthCheckFailureThreshold defines consecutive failures before BGP route withdrawal
+	controlPlaneHealthCheckFailureThreshold = "control_plane_health_check_failure_threshold"
+	// controlPlaneHealthCheckCAPath defines the path to a CA certificate for control-plane health check TLS verification
+	controlPlaneHealthCheckCAPath = "control_plane_health_check_ca_path"
 
 	// zebraEnable defines if Zebra integraton should be enabled
 	zebraEnable = "zebra_enable"
@@ -149,6 +181,9 @@ const (
 	// vipCleanRoutingTable - defines if routing table will be cleaned of redundant routes on kube-vip's start
 	vipCleanRoutingTable = "vip_cleanroutingtable" //nolint
 
+	// vipSkipDAD - defines if Duplicate Address Detection is skipped when adding the VIP address (IFA_F_NODAD)
+	vipSkipDAD = "vip_skipdad" //nolint
+
 	// cpNamespace defines the namespace the control plane pods will run in
 	cpNamespace = "cp_namespace"
 
@@ -209,6 +244,12 @@ const (
 	// dnsMode defines mode that DNS lookup will be performed with (first, ipv4, ipv6, dual)
 	dnsMode = "dns_mode"
 
+	// dhcpMode defines mode that DHCP lookup will be performed with (ipv4, ipv6, dual)
+	dhcpMode = "dhcp_mode"
+
+	// dhcpBackoffAttempts defines how many times DHCP client will try to obtain an IP address
+	dhcpBackoffAttempts = "dhcp_backoff_attempts"
+
 	// disableServiceUpdates disables service updating
 	disableServiceUpdates = "disable_service_updates"
 
@@ -234,4 +275,10 @@ const (
 
 	// egressClean enables egress cleaning on kube-vip's start
 	egressClean = "egress_clean"
+
+	// configFile defines the path to a JSON/YAML configuration file
+	configFile = "config_file"
+
+	// debounceTime defines what time should the event debouncer wait for events
+	debounceTime = "debounce_time"
 )

pkg/kubevip/config_file_test.go

@@ -0,0 +1,569 @@
+package kubevip
+
+import (
+	"os"
+	"path/filepath"
+	"testing"
+)
+
+func TestLoadConfigFromFile(t *testing.T) {
+	// Create temporary directory for test files
+	tmpDir, err := os.MkdirTemp("", "kube-vip-config-test")
+	if err != nil {
+		t.Fatalf("Failed to create temp dir: %v", err)
+	}
+	defer os.RemoveAll(tmpDir)
+
+	tests := []struct {
+		name           string
+		filename       string
+		content        string
+		expectedConfig *Config
+		wantErr        bool
+		errContains    string
+	}{
+		{
+			name:     "Valid YAML config",
+			filename: "config.yaml",
+			content: `
+logging: 2
+enableARP: true
+enableControlPlane: true
+enableServices: true
+address: "192.168.1.100"
+port: 6443
+interface: "eth0"
+namespace: "kube-system"
+instanceName: "release_a"
+vipSubnet: "192.168.1.0/24"
+leaseName: "test-lease"
+leaseDuration: 15
+renewDeadline: 10
+retryPeriod: 2
+prometheusHTTPServer: ":2112"
+`,
+			expectedConfig: &Config{
+				Logging:              2,
+				EnableARP:            true,
+				EnableControlPlane:   true,
+				EnableServices:       true,
+				Address:              "192.168.1.100",
+				Port:                 6443,
+				Interface:            "eth0",
+				Namespace:            "kube-system",
+				InstanceName:         "release_a",
+				VIPSubnet:            "192.168.1.0/24",
+				PrometheusHTTPServer: ":2112",
+				KubernetesLeaderElection: KubernetesLeaderElection{
+					LeaseName:     "test-lease",
+					LeaseDuration: 15,
+					RenewDeadline: 10,
+					RetryPeriod:   2,
+				},
+			},
+			wantErr: false,
+		},
+		{
+			name:     "Valid JSON config",
+			filename: "config.json",
+			content: `{
+  "logging": 3,
+  "enableBGP": true,
+  "enableServices": true,
+  "address": "10.0.0.100",
+  "port": 8443,
+  "interface": "ens192",
+  "namespace": "kube-system",
+  "loadBalancers": [
+    {
+      "name": "control-plane",
+      "ports": [
+        {
+          "type": "TCP",
+          "port": 6443
+        }
+      ],
+      "bindToVip": true,
+      "forwardingMethod": "local"
+    }
+  ]
+}`,
+			expectedConfig: &Config{
+				Logging:        3,
+				EnableBGP:      true,
+				EnableServices: true,
+				Address:        "10.0.0.100",
+				Port:           8443,
+				Interface:      "ens192",
+				Namespace:      "kube-system",
+				LoadBalancers: []LoadBalancer{
+					{
+						Name: "control-plane",
+						Ports: []Port{
+							{
+								Type: "TCP",
+								Port: 6443,
+							},
+						},
+						BindToVip:        true,
+						ForwardingMethod: "local",
+					},
+				},
+			},
+			wantErr: false,
+		},
+		{
+			name:     "Complex BGP config",
+			filename: "bgp-config.yaml",
+			content: `
+enableBGP: true
+bgpConfig:
+  routerID: "192.168.1.1"
+  as: 65000
+  sourceIF: "eth0"
+  holdTime: 60
+  keepaliveInterval: 20
+  peers:
+    - address: "192.168.1.2"
+      as: 65001
+      port: 179
+      multiHop: false
+    - address: "192.168.1.3"
+      as: 65002
+      port: 179
+      multiHop: true
+`,
+			expectedConfig: &Config{
+				EnableBGP: true,
+				BGPConfig: BGPConfig{
+					RouterID:          "192.168.1.1",
+					AS:                65000,
+					SourceIF:          "eth0",
+					HoldTime:          60,
+					KeepaliveInterval: 20,
+					Peers: []BGPPeer{
+						{
+							Address:  "192.168.1.2",
+							AS:       65001,
+							Port:     179,
+							MultiHop: false,
+						},
+						{
+							Address:  "192.168.1.3",
+							AS:       65002,
+							Port:     179,
+							MultiHop: true,
+						},
+					},
+				},
+			},
+			wantErr: false,
+		},
+		{
+			name:        "Invalid JSON",
+			filename:    "invalid.json",
+			content:     `{"logging": 2, "invalid": }`,
+			wantErr:     true,
+			errContains: "failed to parse JSON config file",
+		},
+		{
+			name:        "Invalid YAML",
+			filename:    "invalid.yaml",
+			content:     "logging: 2\ninvalid: [unclosed",
+			wantErr:     true,
+			errContains: "failed to parse YAML config file",
+		},
+		{
+			name:        "Unsupported format",
+			filename:    "config.txt",
+			content:     "logging=2",
+			wantErr:     true,
+			errContains: "unsupported config file format",
+		},
+		{
+			name:        "Empty path",
+			filename:    "",
+			content:     "",
+			wantErr:     true,
+			errContains: "config file path is empty",
+		},
+	}
+
+	for _, tt := range tests {
+		t.Run(tt.name, func(t *testing.T) {
+			var filePath string
+			if tt.filename != "" {
+				filePath = filepath.Join(tmpDir, tt.filename)
+				if err := os.WriteFile(filePath, []byte(tt.content), 0600); err != nil {
+					t.Fatalf("Failed to write test file: %v", err)
+				}
+			}
+
+			config, err := LoadConfigFromFile(filePath)
+
+			if tt.wantErr {
+				if err == nil {
+					t.Errorf("LoadConfigFromFile() expected error, got nil")
+					return
+				}
+				if tt.errContains != "" && !containsString(err.Error(), tt.errContains) {
+					t.Errorf("LoadConfigFromFile() error = %v, expected to contain %v", err, tt.errContains)
+				}
+				return
+			}
+
+			if err != nil {
+				t.Errorf("LoadConfigFromFile() unexpected error = %v", err)
+				return
+			}
+
+			if config == nil {
+				t.Errorf("LoadConfigFromFile() returned nil config")
+				return
+			}
+
+			// Compare key fields
+			if config.Logging != tt.expectedConfig.Logging {
+				t.Errorf("Logging = %v, expected %v", config.Logging, tt.expectedConfig.Logging)
+			}
+			if config.EnableARP != tt.expectedConfig.EnableARP {
+				t.Errorf("EnableARP = %v, expected %v", config.EnableARP, tt.expectedConfig.EnableARP)
+			}
+			if config.EnableBGP != tt.expectedConfig.EnableBGP {
+				t.Errorf("EnableBGP = %v, expected %v", config.EnableBGP, tt.expectedConfig.EnableBGP)
+			}
+			if config.Address != tt.expectedConfig.Address {
+				t.Errorf("Address = %v, expected %v", config.Address, tt.expectedConfig.Address)
+			}
+			if config.Port != tt.expectedConfig.Port {
+				t.Errorf("Port = %v, expected %v", config.Port, tt.expectedConfig.Port)
+			}
+			if config.Interface != tt.expectedConfig.Interface {
+				t.Errorf("Interface = %v, expected %v", config.Interface, tt.expectedConfig.Interface)
+			}
+			if config.InstanceName != tt.expectedConfig.InstanceName {
+				t.Errorf("InstanceName = %v, expected %v", config.InstanceName, tt.expectedConfig.InstanceName)
+			}
+
+			// Test BGP config if present
+			if tt.expectedConfig.EnableBGP {
+				if config.BGPConfig.RouterID != tt.expectedConfig.BGPConfig.RouterID {
+					t.Errorf("BGPConfig.RouterID = %v, expected %v", config.BGPConfig.RouterID, tt.expectedConfig.BGPConfig.RouterID)
+				}
+				if config.BGPConfig.AS != tt.expectedConfig.BGPConfig.AS {
+					t.Errorf("BGPConfig.AS = %v, expected %v", config.BGPConfig.AS, tt.expectedConfig.BGPConfig.AS)
+				}
+				if len(config.BGPConfig.Peers) != len(tt.expectedConfig.BGPConfig.Peers) {
+					t.Errorf("BGPConfig.Peers length = %v, expected %v", len(config.BGPConfig.Peers), len(tt.expectedConfig.BGPConfig.Peers))
+				}
+			}
+
+			// Test LoadBalancers if present
+			if len(tt.expectedConfig.LoadBalancers) > 0 {
+				if len(config.LoadBalancers) != len(tt.expectedConfig.LoadBalancers) {
+					t.Errorf("LoadBalancers length = %v, expected %v", len(config.LoadBalancers), len(tt.expectedConfig.LoadBalancers))
+				}
+			}
+		})
+	}
+}
+
+func TestMergeConfigFromFile(t *testing.T) {
+	// Create temporary directory for test files
+	tmpDir, err := os.MkdirTemp("", "kube-vip-merge-test")
+	if err != nil {
+		t.Fatalf("Failed to create temp dir: %v", err)
+	}
+	defer os.RemoveAll(tmpDir)
+
+	// Create test config file
+	configFile := filepath.Join(tmpDir, "test-config.yaml")
+	configContent := `
+logging: 3
+enableARP: true
+enableServices: true
+address: "192.168.1.200"
+port: 6443
+interface: "eth1"
+namespace: "test-namespace"
+leaseName: "file-lease"
+leaseDuration: 20
+prometheusHTTPServer: ":3000"
+`
+	if err := os.WriteFile(configFile, []byte(configContent), 0600); err != nil {
+		t.Fatalf("Failed to write test config file: %v", err)
+	}
+
+	tests := []struct {
+		name           string
+		baseConfig     *Config
+		configFilePath string
+		expected       *Config
+		wantErr        bool
+	}{
+		{
+			name:           "Merge with empty base config",
+			baseConfig:     &Config{},
+			configFilePath: configFile,
+			expected: &Config{
+				Logging:              3,
+				EnableARP:            true,
+				EnableServices:       true,
+				Address:              "192.168.1.200",
+				Port:                 6443,
+				Interface:            "eth1",
+				Namespace:            "test-namespace",
+				PrometheusHTTPServer: ":3000",
+				KubernetesLeaderElection: KubernetesLeaderElection{
+					LeaseName:     "file-lease",
+					LeaseDuration: 20,
+				},
+			},
+			wantErr: false,
+		},
+		{
+			name: "Merge respects existing values (priority test)",
+			baseConfig: &Config{
+				Logging:   5,      // Should not be overridden
+				Port:      8443,   // Should not be overridden
+				Interface: "eth0", // Should not be overridden
+			},
+			configFilePath: configFile,
+			expected: &Config{
+				Logging:              5,                // From base (higher priority)
+				EnableARP:            true,             // From file
+				EnableServices:       true,             // From file
+				Address:              "192.168.1.200",  // From file
+				Port:                 8443,             // From base (higher priority)
+				Interface:            "eth0",           // From base (higher priority)
+				Namespace:            "test-namespace", // From file
+				PrometheusHTTPServer: ":3000",          // From file
+				KubernetesLeaderElection: KubernetesLeaderElection{
+					LeaseName:     "file-lease", // From file
+					LeaseDuration: 20,           // From file
+				},
+			},
+			wantErr: false,
+		},
+		{
+			name: "Empty config file path",
+			baseConfig: &Config{
+				Logging: 1,
+			},
+			configFilePath: "",
+			expected: &Config{
+				Logging: 1, // Unchanged
+			},
+			wantErr: false,
+		},
+		{
+			name: "Non-existent config file",
+			baseConfig: &Config{
+				Logging: 1,
+			},
+			configFilePath: "/non/existent/file.yaml",
+			expected: &Config{
+				Logging: 1,
+			},
+			wantErr: true,
+		},
+	}
+
+	for _, tt := range tests {
+		t.Run(tt.name, func(t *testing.T) {
+			err := MergeConfigFromFile(tt.baseConfig, tt.configFilePath)
+
+			if tt.wantErr {
+				if err == nil {
+					t.Errorf("MergeConfigFromFile() expected error, got nil")
+				}
+				return
+			}
+
+			if err != nil {
+				t.Errorf("MergeConfigFromFile() unexpected error = %v", err)
+				return
+			}
+
+			// Compare key fields
+			if tt.baseConfig.Logging != tt.expected.Logging {
+				t.Errorf("Logging = %v, expected %v", tt.baseConfig.Logging, tt.expected.Logging)
+			}
+			if tt.baseConfig.EnableARP != tt.expected.EnableARP {
+				t.Errorf("EnableARP = %v, expected %v", tt.baseConfig.EnableARP, tt.expected.EnableARP)
+			}
+			if tt.baseConfig.Address != tt.expected.Address {
+				t.Errorf("Address = %v, expected %v", tt.baseConfig.Address, tt.expected.Address)
+			}
+			if tt.baseConfig.Port != tt.expected.Port {
+				t.Errorf("Port = %v, expected %v", tt.baseConfig.Port, tt.expected.Port)
+			}
+			if tt.baseConfig.Interface != tt.expected.Interface {
+				t.Errorf("Interface = %v, expected %v", tt.baseConfig.Interface, tt.expected.Interface)
+			}
+			if tt.baseConfig.Namespace != tt.expected.Namespace {
+				t.Errorf("Namespace = %v, expected %v", tt.baseConfig.Namespace, tt.expected.Namespace)
+			}
+			if tt.baseConfig.PrometheusHTTPServer != tt.expected.PrometheusHTTPServer {
+				t.Errorf("PrometheusHTTPServer = %v, expected %v", tt.baseConfig.PrometheusHTTPServer, tt.expected.PrometheusHTTPServer)
+			}
+		})
+	}
+}
+
+func TestMergeConfigValues(t *testing.T) {
+	tests := []struct {
+		name         string
+		baseConfig   *Config
+		fileConfig   *Config
+		expectedBase *Config
+	}{
+		{
+			name: "Merge basic configuration",
+			baseConfig: &Config{
+				Logging: 5, // Should not be overridden
+				Port:    0, // Should be overridden
+			},
+			fileConfig: &Config{
+				Logging:      2,
+				Port:         6443,
+				Interface:    "eth0",
+				Address:      "192.168.1.100",
+				InstanceName: "release_a",
+			},
+			expectedBase: &Config{
+				Logging:      5,               // From base (non-zero)
+				Port:         6443,            // From file (base was zero)
+				Interface:    "eth0",          // From file (base was empty)
+				Address:      "192.168.1.100", // From file (base was empty)
+				InstanceName: "release_a",     // From file (base was empty)
+			},
+		},
+		{
+			name: "Merge boolean flags",
+			baseConfig: &Config{
+				EnableARP: true, // Should not be overridden
+			},
+			fileConfig: &Config{
+				EnableARP:       false, // Should not override true
+				EnableBGP:       true,  // Should be set
+				EnableServices:  true,  // Should be set
+				EnableWireguard: false, // Should not be set (false doesn't override false)
+			},
+			expectedBase: &Config{
+				EnableARP:       true,  // From base (true has priority)
+				EnableBGP:       true,  // From file
+				EnableServices:  true,  // From file
+				EnableWireguard: false, // Remains false
+			},
+		},
+		{
+			name: "Merge BGP configuration",
+			baseConfig: &Config{
+				BGPConfig: BGPConfig{
+					RouterID: "1.1.1.1", // Should not be overridden
+				},
+			},
+			fileConfig: &Config{
+				BGPConfig: BGPConfig{
+					RouterID:          "2.2.2.2", // Should not override
+					AS:                65000,     // Should be set
+					SourceIF:          "eth0",    // Should be set
+					HoldTime:          30,        // Should be set
+					KeepaliveInterval: 10,        // Should be set
+				},
+			},
+			expectedBase: &Config{
+				BGPConfig: BGPConfig{
+					RouterID:          "1.1.1.1", // From base (non-empty)
+					AS:                65000,     // From file (base was zero)
+					SourceIF:          "eth0",    // From file (base was empty)
+					HoldTime:          30,        // From file (base was zero)
+					KeepaliveInterval: 10,        // From file (base was zero)
+				},
+			},
+		},
+		{
+			name: "Merge leader election configuration",
+			baseConfig: &Config{
+				KubernetesLeaderElection: KubernetesLeaderElection{
+					LeaseName: "base-lease", // Should not be overridden
+				},
+			},
+			fileConfig: &Config{
+				KubernetesLeaderElection: KubernetesLeaderElection{
+					LeaseName:     "file-lease", // Should not override
+					LeaseDuration: 15,           // Should be set
+					RenewDeadline: 10,           // Should be set
+					RetryPeriod:   2,            // Should be set
+				},
+			},
+			expectedBase: &Config{
+				KubernetesLeaderElection: KubernetesLeaderElection{
+					LeaseName:     "base-lease", // From base (non-empty)
+					LeaseDuration: 15,           // From file (base was zero)
+					RenewDeadline: 10,           // From file (base was zero)
+					RetryPeriod:   2,            // From file (base was zero)
+				},
+			},
+		},
+	}
+
+	for _, tt := range tests {
+		t.Run(tt.name, func(t *testing.T) {
+			mergeConfigValues(tt.baseConfig, tt.fileConfig)
+
+			// Compare results
+			if tt.baseConfig.Logging != tt.expectedBase.Logging {
+				t.Errorf("Logging = %v, expected %v", tt.baseConfig.Logging, tt.expectedBase.Logging)
+			}
+			if tt.baseConfig.Port != tt.expectedBase.Port {
+				t.Errorf("Port = %v, expected %v", tt.baseConfig.Port, tt.expectedBase.Port)
+			}
+			if tt.baseConfig.Interface != tt.expectedBase.Interface {
+				t.Errorf("Interface = %v, expected %v", tt.baseConfig.Interface, tt.expectedBase.Interface)
+			}
+			if tt.baseConfig.InstanceName != tt.expectedBase.InstanceName {
+				t.Errorf("InstanceName = %v, expected %v", tt.baseConfig.InstanceName, tt.expectedBase.InstanceName)
+			}
+			if tt.baseConfig.EnableARP != tt.expectedBase.EnableARP {
+				t.Errorf("EnableARP = %v, expected %v", tt.baseConfig.EnableARP, tt.expectedBase.EnableARP)
+			}
+			if tt.baseConfig.EnableBGP != tt.expectedBase.EnableBGP {
+				t.Errorf("EnableBGP = %v, expected %v", tt.baseConfig.EnableBGP, tt.expectedBase.EnableBGP)
+			}
+			if tt.baseConfig.BGPConfig.RouterID != tt.expectedBase.BGPConfig.RouterID {
+				t.Errorf("BGPConfig.RouterID = %v, expected %v", tt.baseConfig.BGPConfig.RouterID, tt.expectedBase.BGPConfig.RouterID)
+			}
+			if tt.baseConfig.BGPConfig.AS != tt.expectedBase.BGPConfig.AS {
+				t.Errorf("BGPConfig.AS = %v, expected %v", tt.baseConfig.BGPConfig.AS, tt.expectedBase.BGPConfig.AS)
+			}
+			if tt.baseConfig.KubernetesLeaderElection.LeaseName != tt.expectedBase.KubernetesLeaderElection.LeaseName {
+				t.Errorf("KubernetesLeaderElection.LeaseName = %v, expected %v", tt.baseConfig.KubernetesLeaderElection.LeaseName, tt.expectedBase.KubernetesLeaderElection.LeaseName)
+			}
+		})
+	}
+}
+
+func TestLoadConfigFromFile_FileNotExists(t *testing.T) {
+	_, err := LoadConfigFromFile("/non/existent/path/config.yaml")
+	if err == nil {
+		t.Error("LoadConfigFromFile() expected error for non-existent file, got nil")
+	}
+	if !containsString(err.Error(), "config file does not exist") {
+		t.Errorf("LoadConfigFromFile() error = %v, expected to contain 'config file does not exist'", err)
+	}
+}
+
+// Helper function to check if a string contains a substring
+func containsString(str, substr string) bool {
+	return len(str) >= len(substr) && (str == substr || len(substr) == 0 ||
+		(len(substr) > 0 && func() bool {
+			for i := 0; i <= len(str)-len(substr); i++ {
+				if str[i:i+len(substr)] == substr {
+					return true
+				}
+			}
+			return false
+		}()))
+}

pkg/kubevip/config_generator.go

@@ -5,6 +5,8 @@ import (
 	"log"
 	"strconv"
 
+	"github.com/google/go-containerregistry/pkg/name"
+	"github.com/kube-vip/kube-vip/pkg/debouncer"
 	appv1 "k8s.io/api/apps/v1"
 	corev1 "k8s.io/api/core/v1"
 	metav1 "k8s.io/apimachinery/pkg/apis/meta/v1"
@@ -13,7 +15,6 @@ import (
 	applyCoreV1 "k8s.io/client-go/applyconfigurations/core/v1"
 	applyMetaV1 "k8s.io/client-go/applyconfigurations/meta/v1"
 	applyRbacV1 "k8s.io/client-go/applyconfigurations/rbac/v1"
-
 	"sigs.k8s.io/yaml"
 )
 
@@ -167,7 +168,11 @@ func GenerateRoleBinding(rolebinding bool, saCfg *applyCoreV1.ServiceAccountAppl
 }
 
 // generatePodSpec will take a kube-vip config and generate a Pod spec
-func generatePodSpec(c *Config, imageVersion string, inCluster bool) *corev1.Pod {
+func generatePodSpec(c *Config, image, imageVersion string, inCluster bool) (*corev1.Pod, error) {
+	imageRef, err := name.NewTag(image, name.WeakValidation, name.WithDefaultTag(imageVersion))
+	if err != nil {
+		return nil, fmt.Errorf("cannot parse %q: %w", image, err)
+	}
 	command := "manager"
 
 	// Determine where the pods should be living (for multi-tenancy)
@@ -197,6 +202,12 @@ func generatePodSpec(c *Config, imageVersion string, inCluster bool) *corev1.Pod
 			},
 		},
 	}
+	if c.InstanceName != "" {
+		newEnvironment = append(newEnvironment, corev1.EnvVar{
+			Name:  instanceName,
+			Value: c.InstanceName,
+		})
+	}
 
 	// If we're specifically saying which interface to use then add it to the manifest
 	if c.Interface != "" {
@@ -228,6 +239,17 @@ func generatePodSpec(c *Config, imageVersion string, inCluster bool) *corev1.Pod
 		newEnvironment = append(newEnvironment, svcInterface...)
 	}
 
+	// Tolerate a down interface
+	if c.AllowInterfaceNotUp {
+		allowIface := []corev1.EnvVar{
+			{
+				Name:  vipAllowInterfaceNotUp,
+				Value: strconv.FormatBool(c.AllowInterfaceNotUp),
+			},
+		}
+		newEnvironment = append(newEnvironment, allowIface...)
+	}
+
 	// If a subnet is required for the VIP
 	if c.VIPSubnet != "" {
 		// build environment variables
@@ -251,6 +273,28 @@ func generatePodSpec(c *Config, imageVersion string, inCluster bool) *corev1.Pod
 		newEnvironment = append(newEnvironment, dnsModeSelector...)
 	}
 
+	if c.DHCPMode != "" {
+		// build environment variables
+		dhcpModeSelector := []corev1.EnvVar{
+			{
+				Name:  dhcpMode,
+				Value: c.DHCPMode,
+			},
+		}
+		newEnvironment = append(newEnvironment, dhcpModeSelector...)
+	}
+
+	if c.DHCPBackoffAttempts != DefaultDHCPBackoffAttempts {
+		// build environment variables
+		dhcpBackoff := []corev1.EnvVar{
+			{
+				Name:  dhcpBackoffAttempts,
+				Value: strconv.FormatUint(uint64(c.DHCPBackoffAttempts), 10),
+			},
+		}
+		newEnvironment = append(newEnvironment, dhcpBackoff...)
+	}
+
 	// If we're doing the hybrid mode
 	if c.EnableControlPlane {
 		cp := []corev1.EnvVar{
@@ -363,6 +407,23 @@ func generatePodSpec(c *Config, imageVersion string, inCluster bool) *corev1.Pod
 		newEnvironment = append(newEnvironment, leaderElection...)
 	}
 
+	if c.LoseLeadership {
+		loseLeadership := []corev1.EnvVar{
+			{
+				Name:  vipLoseLeadership,
+				Value: strconv.FormatBool(c.LoseLeadership),
+			},
+		}
+
+		if c.LoseLeadershipTimeoutSeconds > 0 {
+			loseLeadership = append(loseLeadership, corev1.EnvVar{
+				Name:  vipLoseLeadership,
+				Value: fmt.Sprintf("%d", c.LoseLeadershipTimeoutSeconds),
+			})
+		}
+		newEnvironment = append(newEnvironment, loseLeadership...)
+	}
+
 	// If we're enabling node labeling on leader election
 	if c.EnableNodeLabeling {
 		EnableNodeLabeling := []corev1.EnvVar{
@@ -415,6 +476,12 @@ func generatePodSpec(c *Config, imageVersion string, inCluster bool) *corev1.Pod
 				Value: strconv.FormatBool(c.EnableBGP),
 			},
 		}
+		if c.BGPAttachIPToInterface {
+			bgp = append(bgp, corev1.EnvVar{
+				Name:  bgpAttachIPToInterface,
+				Value: strconv.FormatBool(c.BGPAttachIPToInterface),
+			})
+		}
 		newEnvironment = append(newEnvironment, bgp...)
 	}
 
@@ -481,6 +548,40 @@ func generatePodSpec(c *Config, imageVersion string, inCluster bool) *corev1.Pod
 
 	}
 
+	if c.ControlPlaneHealthCheck.Address != "" {
+		healthCheckVars := []corev1.EnvVar{
+			{
+				Name:  controlPlaneHealthCheckAddress,
+				Value: c.ControlPlaneHealthCheck.Address,
+			},
+		}
+		if c.ControlPlaneHealthCheck.PeriodSeconds > 0 {
+			healthCheckVars = append(healthCheckVars, corev1.EnvVar{
+				Name:  controlPlaneHealthCheckPeriodSeconds,
+				Value: fmt.Sprintf("%d", c.ControlPlaneHealthCheck.PeriodSeconds),
+			})
+		}
+		if c.ControlPlaneHealthCheck.TimeoutSeconds > 0 {
+			healthCheckVars = append(healthCheckVars, corev1.EnvVar{
+				Name:  controlPlaneHealthCheckTimeoutSeconds,
+				Value: fmt.Sprintf("%d", c.ControlPlaneHealthCheck.TimeoutSeconds),
+			})
+		}
+		if c.ControlPlaneHealthCheck.FailureThreshold > 0 {
+			healthCheckVars = append(healthCheckVars, corev1.EnvVar{
+				Name:  controlPlaneHealthCheckFailureThreshold,
+				Value: fmt.Sprintf("%d", c.ControlPlaneHealthCheck.FailureThreshold),
+			})
+		}
+		if c.ControlPlaneHealthCheck.CAPath != "" {
+			healthCheckVars = append(healthCheckVars, corev1.EnvVar{
+				Name:  controlPlaneHealthCheckCAPath,
+				Value: c.ControlPlaneHealthCheck.CAPath,
+			})
+		}
+		newEnvironment = append(newEnvironment, healthCheckVars...)
+	}
+
 	// If the load-balancer is enabled then add the configuration to the manifest
 	if c.EnableLoadBalancer {
 		lb := []corev1.EnvVar{
@@ -561,7 +662,7 @@ func generatePodSpec(c *Config, imageVersion string, inCluster bool) *corev1.Pod
 
 	var securityContext *corev1.SecurityContext
 	if c.LoadBalancerForwardingMethod == "masquerade" {
-		var privileged = true
+		privileged := true
 		securityContext = &corev1.SecurityContext{
 			Privileged: &privileged,
 		}
@@ -579,6 +680,25 @@ func generatePodSpec(c *Config, imageVersion string, inCluster bool) *corev1.Pod
 		}
 	}
 
+	if c.PreserveVIPOnLeadershipLoss {
+		preserveVIPOnLeadershipLoss := []corev1.EnvVar{
+			{
+				Name:  vipPreserveOnLeadershipLoss,
+				Value: strconv.FormatBool(c.PreserveVIPOnLeadershipLoss),
+			},
+		}
+		newEnvironment = append(newEnvironment, preserveVIPOnLeadershipLoss...)
+	}
+
+	if c.DebounceTime != debouncer.DefaultTime {
+		debTime := corev1.EnvVar{
+			Name:  debounceTime,
+			Value: c.DebounceTime,
+		}
+
+		newEnvironment = append(newEnvironment, debTime)
+	}
+
 	newManifest := &corev1.Pod{
 		TypeMeta: metav1.TypeMeta{
 			Kind:       "Pod",
@@ -592,7 +712,7 @@ func generatePodSpec(c *Config, imageVersion string, inCluster bool) *corev1.Pod
 			Containers: []corev1.Container{
 				{
 					Name:            "kube-vip",
-					Image:           fmt.Sprintf("ghcr.io/kube-vip/kube-vip:%s", imageVersion),
+					Image:           imageRef.Name(),
 					ImagePullPolicy: corev1.PullIfNotPresent,
 					SecurityContext: securityContext,
 					Args: []string{
@@ -633,18 +753,24 @@ func generatePodSpec(c *Config, imageVersion string, inCluster bool) *corev1.Pod
 		newManifest.Spec.HostAliases = append(newManifest.Spec.HostAliases, hostAlias)
 	}
 
-	return newManifest
+	return newManifest, nil
 }
 
 // GeneratePodManifestFromConfig will take a kube-vip config and generate a manifest
-func GeneratePodManifestFromConfig(c *Config, imageVersion string, inCluster bool) string {
-	newManifest := generatePodSpec(c, imageVersion, inCluster)
-	b, _ := yaml.Marshal(newManifest)
-	return string(b)
+func GeneratePodManifestFromConfig(c *Config, image, imageVersion string, inCluster bool) (string, error) {
+	newManifest, err := generatePodSpec(c, image, imageVersion, inCluster)
+	if err != nil {
+		return "", err
+	}
+	b, err := yaml.Marshal(newManifest)
+	if err != nil {
+		return "", fmt.Errorf("failed to marshal manifest: %w", err)
+	}
+	return string(b), nil
 }
 
 // GenerateDaemonsetManifestFromConfig will take a kube-vip config and generate a manifest
-func GenerateDaemonsetManifestFromConfig(c *Config, imageVersion string, inCluster, taint bool) string {
+func GenerateDaemonsetManifestFromConfig(c *Config, image, imageVersion string, inCluster, taint bool) (string, error) {
 	// Determine where the pod should be deployed
 	var namespace string
 	if c.ServiceNamespace != "" {
@@ -653,7 +779,11 @@ func GenerateDaemonsetManifestFromConfig(c *Config, imageVersion string, inClust
 		namespace = metav1.NamespaceSystem
 	}
 
-	podSpec := generatePodSpec(c, imageVersion, inCluster).Spec
+	pod, err := generatePodSpec(c, image, imageVersion, inCluster)
+	if err != nil {
+		return "", err
+	}
+
 	newManifest := &appv1.DaemonSet{
 		TypeMeta: metav1.TypeMeta{
 			Kind:       "DaemonSet",
@@ -680,7 +810,7 @@ func GenerateDaemonsetManifestFromConfig(c *Config, imageVersion string, inClust
 						"app.kubernetes.io/version": imageVersion,
 					},
 				},
-				Spec: podSpec,
+				Spec: pod.Spec,
 			},
 		},
 	}
@@ -730,5 +860,5 @@ func GenerateDaemonsetManifestFromConfig(c *Config, imageVersion string, inClust
 	delete(m, "status")
 
 	b, _ = yaml.Marshal(m)
-	return string(b)
+	return string(b), nil
 }

pkg/kubevip/config_generator_test.go

@@ -1,6 +1,10 @@
 package kubevip
 
-import "testing"
+import (
+	"os"
+	"strings"
+	"testing"
+)
 
 func TestParseEnvironment(t *testing.T) {
 
@@ -9,8 +13,8 @@ func TestParseEnvironment(t *testing.T) {
 		c       *Config
 		wantErr bool
 	}{
-		{"", nil, false},
-		{"", &Config{Interface: "eth0", ServicesInterface: "eth1"}, false},
+		{"nil config", nil, false},
+		{"basic config", &Config{Interface: "eth0", ServicesInterface: "eth1"}, false},
 	}
 	for _, tt := range tests {
 		t.Logf("%v", tt.c)
@@ -21,3 +25,146 @@ func TestParseEnvironment(t *testing.T) {
 		})
 	}
 }
+
+func TestParseEnvironmentInstanceName(t *testing.T) {
+	tests := []struct {
+		name      string
+		lowercase string
+		uppercase string
+		want      string
+	}{
+		{name: "lowercase", lowercase: "release_a", want: "release_a"},
+		{name: "uppercase fallback", uppercase: "release_b", want: "release_b"},
+		{name: "lowercase takes precedence", lowercase: "release_a", uppercase: "release_b", want: "release_a"},
+	}
+
+	for _, tt := range tests {
+		t.Run(tt.name, func(t *testing.T) {
+			t.Setenv(instanceName, tt.lowercase)
+			t.Setenv(strings.ToUpper(instanceName), tt.uppercase)
+
+			config := &Config{}
+			if err := ParseEnvironment(config); err != nil {
+				t.Fatalf("ParseEnvironment() error = %v", err)
+			}
+			if config.InstanceName != tt.want {
+				t.Fatalf("InstanceName = %q, want %q", config.InstanceName, tt.want)
+			}
+		})
+	}
+}
+
+func TestParseEnvironmentBGPAttachIPToInterface(t *testing.T) {
+	t.Setenv(bgpAttachIPToInterface, "true")
+
+	config := &Config{}
+	if err := ParseEnvironment(config); err != nil {
+		t.Fatalf("ParseEnvironment() error = %v", err)
+	}
+	if !config.BGPAttachIPToInterface {
+		t.Fatal("BGPAttachIPToInterface = false, want true")
+	}
+}
+
+func TestGeneratePodSpecBGPAttachIPToInterface(t *testing.T) {
+	pod, err := generatePodSpec(&Config{
+		EnableBGP:              true,
+		BGPAttachIPToInterface: true,
+	}, "ghcr.io/kube-vip/kube-vip", "v0.0.0", true)
+	if err != nil {
+		t.Fatalf("generatePodSpec() error = %v", err)
+	}
+
+	for _, env := range pod.Spec.Containers[0].Env {
+		if env.Name == bgpAttachIPToInterface && env.Value == "true" {
+			return
+		}
+	}
+	t.Fatalf("%s=true is missing from generated pod environment", bgpAttachIPToInterface)
+}
+
+func TestGeneratePodSpecInstanceName(t *testing.T) {
+	tests := []struct {
+		name         string
+		instanceName string
+		wantPresent  bool
+	}{
+		{name: "configured", instanceName: "release_a", wantPresent: true},
+		{name: "empty", wantPresent: false},
+	}
+
+	for _, tt := range tests {
+		t.Run(tt.name, func(t *testing.T) {
+			pod, err := generatePodSpec(&Config{InstanceName: tt.instanceName}, "ghcr.io/kube-vip/kube-vip", "v0.0.0", true)
+			if err != nil {
+				t.Fatalf("generatePodSpec() error = %v", err)
+			}
+
+			var value string
+			found := false
+			for _, env := range pod.Spec.Containers[0].Env {
+				if env.Name == instanceName {
+					found = true
+					value = env.Value
+					break
+				}
+			}
+			if found != tt.wantPresent {
+				t.Fatalf("instance_name present = %t, want %t", found, tt.wantPresent)
+			}
+			if found && value != tt.instanceName {
+				t.Fatalf("instance_name = %q, want %q", value, tt.instanceName)
+			}
+		})
+	}
+}
+
+func TestParseEnvironmentConfigFile(t *testing.T) {
+	// Save original environment
+	originalConfigFile := os.Getenv("config_file")
+	defer func() {
+		if originalConfigFile != "" {
+			os.Setenv("config_file", originalConfigFile)
+		} else {
+			os.Unsetenv("config_file")
+		}
+	}()
+
+	tests := []struct {
+		name           string
+		envValue       string
+		expectedConfig string
+	}{
+		{
+			name:           "config_file environment variable set",
+			envValue:       "/etc/kube-vip/config.yaml",
+			expectedConfig: "/etc/kube-vip/config.yaml",
+		},
+		{
+			name:           "config_file environment variable empty",
+			envValue:       "",
+			expectedConfig: "",
+		},
+	}
+
+	for _, tt := range tests {
+		t.Run(tt.name, func(t *testing.T) {
+			// Set environment variable
+			if tt.envValue != "" {
+				os.Setenv("config_file", tt.envValue)
+			} else {
+				os.Unsetenv("config_file")
+			}
+
+			config := &Config{}
+			err := ParseEnvironment(config)
+			if err != nil {
+				t.Errorf("ParseEnvironment() unexpected error = %v", err)
+			}
+
+			if config.ConfigFile != tt.expectedConfig {
+				t.Errorf("ConfigFile = %v, expected %v", config.ConfigFile, tt.expectedConfig)
+			}
+		})
+	}
+}

pkg/kubevip/config_manager.go

@@ -1,6 +1,7 @@
 package kubevip
 
 import (
+	"errors"
 	"fmt"
 
 	log "log/slog"
@@ -12,16 +13,34 @@ const (
 	Auto = "auto"
 )
 
+var ErrInterfaceNotUp = errors.New("interface is not up")
+
+func (c *Config) CheckSubnetExists() error {
+	if c.VIPSubnet == "" && c.VIP != "" && c.Address == "" {
+		return fmt.Errorf("vip_subnet must be set if using vip_address instead of address environment variable")
+	}
+
+	return nil
+}
+
 func (c *Config) CheckInterface() error {
 	if c.Interface != "" {
 		if err := isValidInterface(c.Interface); err != nil {
-			return fmt.Errorf("%s is not valid interface, reason: %w", c.Interface, err)
+			if errors.Is(err, ErrInterfaceNotUp) && c.AllowInterfaceNotUp {
+				log.Warn("interface is not up, continuing as allowInterfaceNotUp is set", "interface", c.Interface)
+			} else {
+				return fmt.Errorf("%s is not valid interface, reason: %w", c.Interface, err)
+			}
 		}
 	}
 
 	if c.ServicesInterface != "" {
 		if err := isValidInterface(c.ServicesInterface); err != nil {
-			return fmt.Errorf("%s is not valid interface, reason: %w", c.ServicesInterface, err)
+			if errors.Is(err, ErrInterfaceNotUp) && c.AllowInterfaceNotUp {
+				log.Warn("interface is not up, continuing as allowInterfaceNotUp is set", "interface", c.ServicesInterface)
+			} else {
+				return fmt.Errorf("%s is not valid interface, reason: %w", c.ServicesInterface, err)
+			}
 		}
 	}
 
@@ -50,7 +69,7 @@ func isValidInterface(iface string) error {
 			iface,
 		)
 	} else if attrs.OperState != netlink.OperUp {
-		return fmt.Errorf("%s is not up", iface)
+		return fmt.Errorf("%s %w", iface, ErrInterfaceNotUp)
 	}
 
 	return nil

pkg/kubevip/config_manager_test.go

@@ -0,0 +1,56 @@
+package kubevip
+
+import "testing"
+
+func TestCheckSubnetExists(t *testing.T) {
+	tests := []struct {
+		name    string
+		config  Config
+		wantErr bool
+	}{
+		{
+			name: "vip only without subnet returns error",
+			config: Config{
+				VIP: "172.18.0.20",
+			},
+			wantErr: true,
+		},
+		{
+			name: "vip with subnet does not return error",
+			config: Config{
+				VIP:       "172.18.0.20",
+				VIPSubnet: "32",
+			},
+			wantErr: false,
+		},
+		{
+			name: "address only without subnet does not return error",
+			config: Config{
+				Address: "172.18.0.20",
+			},
+			wantErr: false,
+		},
+		{
+			name: "address overrides vip without subnet",
+			config: Config{
+				VIP:     "172.18.0.20",
+				Address: "172.18.0.30",
+			},
+			wantErr: false,
+		},
+		{
+			name:    "empty config does not return error",
+			config:  Config{},
+			wantErr: false,
+		},
+	}
+
+	for _, tt := range tests {
+		t.Run(tt.name, func(t *testing.T) {
+			err := tt.config.CheckSubnetExists()
+			if (err != nil) != tt.wantErr {
+				t.Errorf("CheckSubnetExists() error = %v, wantErr %v", err, tt.wantErr)
+			}
+		})
+	}
+}

pkg/kubevip/config_types.go

@@ -1,9 +1,5 @@
 package kubevip
 
-import (
-	"github.com/kube-vip/kube-vip/pkg/bgp"
-)
-
 // Config defines all of the settings for the Kube-Vip Pod
 type Config struct {
 	// Logging, settings
@@ -15,6 +11,9 @@ type Config struct {
 	// EnableBGP, will use BGP to advertise the VIP address
 	EnableBGP bool `yaml:"enableBGP"`
 
+	// BGPAttachIPToInterface assigns BGP-advertised service VIPs to the configured interface
+	BGPAttachIPToInterface bool `yaml:"bgpAttachIPToInterface"`
+
 	// EnableWireguard, will use wireguard to advertise the VIP address
 	EnableWireguard bool `yaml:"enableWireguard"`
 
@@ -54,6 +53,16 @@ type Config struct {
 	// ArpBroadcastRate, defines how often kube-vip will update the network about updates to the network
 	ArpBroadcastRate int64 `yaml:"arpBroadcastRate"`
 
+	// PreserveVIPOnLeadershipLoss, if true, VIP addresses will remain on interface when leadership is lost (only ARP/NDP broadcasting stops)
+	// If false, VIP addresses are deleted on leadership loss (legacy behavior)
+	PreserveVIPOnLeadershipLoss bool `yaml:"preserveVipOnLeadershipLoss"`
+
+	// LoseLeadership enables leadership loss if VIP interface(physical) is down
+	LoseLeadership bool `yaml:"loseLeadership"`
+
+	// LoseLeadershipTimeoutSeconds defines the timeout after which interface will be considered down. Default is 30s
+	LoseLeadershipTimeoutSeconds int `yaml:"loseLeadershipTimeoutSeconds"`
+
 	// Annotations will define if we're going to wait and lookup configuration from Kubernetes node annotations
 	Annotations string
 
@@ -85,6 +94,9 @@ type Config struct {
 	// Namespace will define which namespace the control plane pods will run in
 	Namespace string `yaml:"namespace"`
 
+	// InstanceName identifies this kube-vip deployment when naming host-global resources.
+	InstanceName string `yaml:"instanceName"`
+
 	// Namespace will define which namespace the control plane pods will run in
 	ServiceNamespace string `yaml:"serviceNamespace"`
 
@@ -106,6 +118,9 @@ type Config struct {
 	// ServicesInterface is the network interface to bind to for services (optional)
 	ServicesInterface string `yaml:"servicesInterface,omitempty"`
 
+	// AllowInterfaceNotUp allows kube-vip to start even when the interface is not up
+	AllowInterfaceNotUp bool `yaml:"allowInterfaceNotUp,omitempty"`
+
 	// EnableLoadBalancer, provides the flexibility to make the load-balancer optional
 	EnableLoadBalancer bool `yaml:"enableLoadBalancer"`
 
@@ -127,11 +142,18 @@ type Config struct {
 	// Clean routing table of redundant routes on start
 	CleanRoutingTable bool `yaml:"cleanRoutingTable"`
 
+	// Skip Duplicate Address Detection when adding the VIP address (IFA_F_NODAD)
+	SkipDAD bool `yaml:"skipDAD"`
+
 	// BGP Configuration
-	BGPConfig     bgp.Config
-	BGPPeerConfig bgp.Peer
+	BGPConfig     BGPConfig
+	BGPPeerConfig BGPPeer
 	BGPPeers      []string
 
+	// ControlPlaneHealthCheck configures HTTP polling of the control plane when using BGP without
+	// leader election. If the health check fails, the BGP route will be withdrawn.
+	ControlPlaneHealthCheck HealthCheck `yaml:"controlPlaneHealthCheck,omitempty"`
+
 	// LoadBalancers are the various services we can load balance over
 	LoadBalancers []LoadBalancer `yaml:"loadBalancers,omitempty"`
 
@@ -146,6 +168,9 @@ type Config struct {
 	// EgressServiceCidr, this contains the service cidr range to ignore
 	EgressServiceCidr string
 
+	// EnableInternalSNAT, this will enable the internal SNAT rule that kube-vip adds to the egress chain
+	EnableInternalSNAT bool
+
 	// EgressWithNftables, this will use the iptables-nftables OVER iptables
 	EgressWithNftables bool
 
@@ -158,6 +183,16 @@ type Config struct {
 	// DNSMode, this will set the mode DSN lookup will be performed (first, ipv4, ipv6, dual)
 	DNSMode string `yaml:"dnsDualStackMode"`
 
+	// IsDualStack reports if service is DualStack.
+	IsDualStack bool
+
+	// RequireDualStack defines if DualStack is required for the service. Based on service's Spec.ipFamilyPolicy field.
+	RequireDualStack bool
+
+	// DNSMode, this will set the mode DHCP lookup will be performed for DDNS (ipv4, ipv6, dual). By default will be the same as DNSMode.
+	// If DNSMode is 'first', IPv4 will be used.
+	DHCPMode string `yaml:"dhcpDualStackMode"`
+
 	// DisableServiceUpdates, if true, kube-vip will only advertise service, but it will not update service's Status.LoadBalancer.Ingress slice
 	DisableServiceUpdates bool `yaml:"disableServiceUpdates"`
 
@@ -186,6 +221,18 @@ type Config struct {
 
 	// EgressClean, enables egress cleaning on Kube-vip's start
 	EgressClean bool `yaml:"egressClean"`
+
+	// ConfigFile defines the path to a JSON/YAML configuration file
+	ConfigFile string `yaml:"configFile"`
+
+	// DHCPBackoffAttempts defines how many times will DHCP client try to obtain address (unlimited when 0)
+	DHCPBackoffAttempts uint `yaml:"dhcpBackoffAttempts"`
+
+	// DebounceTime defines how long will event debouncer wait for the events to arrive
+	DebounceTime string `yaml:"debounceTime"`
+
+	// PerServiceElectionOnDemand will enable kube-vip to handle services with per-service election when annotation is used
+	PerServiceElectionOnDemand bool `yaml:"perServiceElectionOnDemand"`
 }
 
 // KubernetesLeaderElection defines all of the settings for Kubernetes KubernetesLeaderElection
@@ -217,12 +264,29 @@ type Etcd struct {
 	Endpoints      []string
 }
 
+// HealthCheck defines HTTP health-check settings for control-plane polling when using BGP
+// without leader election.
+type HealthCheck struct {
+	// Address is the URL to poll to check the health of the control-plane. If the health
+	// check fails, the BGP route will be withdrawn.
+	Address string `yaml:"address"`
+	// PeriodSeconds is the interval in seconds between health checks.
+	PeriodSeconds int `yaml:"periodSeconds"`
+	// TimeoutSeconds is the timeout per health check request. If a request takes longer
+	// than this timeout, the health check is considered failed.
+	TimeoutSeconds int `yaml:"timeoutSeconds"`
+	// FailureThreshold is the number of consecutive failures before route withdrawal.
+	FailureThreshold int `yaml:"failureThreshold"`
+	// CAPath is the CA certificate path used for TLS verification when Address is an HTTPS URL.
+	CAPath string `yaml:"caPath"`
+}
+
 // LoadBalancer contains the configuration of a load balancing instance
 type LoadBalancer struct {
 	// Name of a LoadBalancer
 	Name string `yaml:"name"`
 
-	//Ports exposed by a LoadBalancer
+	// Ports exposed by a LoadBalancer
 	Ports []Port
 
 	// BindToVip will bind the load balancer port to the VIP itself

pkg/kubevip/config_validation.go

@@ -0,0 +1,73 @@
+package kubevip
+
+import (
+	"fmt"
+	"net/url"
+	"strings"
+)
+
+const (
+	// nftables object names are limited to 255 bytes. Reserve space for the
+	// prefix and address-family suffix added to the instance name.
+	nftablesNameMaxLength     = 255
+	egressNftablesTablePrefix = "kube_vip_"
+	egressNftablesTableSuffix = "_v4"
+	instanceNameMaxLength     = nftablesNameMaxLength - len(egressNftablesTablePrefix) - len(egressNftablesTableSuffix)
+)
+
+// Validate runs configuration checks that are independent of host state.
+// This should be called after all config sources (flags, file, env vars) are merged.
+func (c *Config) Validate() error {
+	if err := validateHealthCheckAddress(c.ControlPlaneHealthCheck.Address); err != nil {
+		return err
+	}
+	if err := validateInstanceName(c.InstanceName); err != nil {
+		return err
+	}
+
+	return nil
+}
+
+func validateInstanceName(name string) error {
+	if name == "" {
+		return nil
+	}
+	if len(name) > instanceNameMaxLength {
+		return fmt.Errorf("instance_name is %d bytes, must not exceed %d bytes so the %q prefix and %q or %q suffix fit within the nftables %d-byte name limit",
+			len(name), instanceNameMaxLength, egressNftablesTablePrefix, "_v4", "_v6", nftablesNameMaxLength)
+	}
+
+	for position, char := range name {
+		if isValidNftablesNameCharacter(char) {
+			continue
+		}
+		return fmt.Errorf("instance_name %q contains invalid character %q at byte %d; only ASCII letters, digits, '.', '-' and '_' are allowed",
+			name, char, position)
+	}
+	return nil
+}
+
+func isValidNftablesNameCharacter(char rune) bool {
+	return char >= 'a' && char <= 'z' ||
+		char >= 'A' && char <= 'Z' ||
+		char >= '0' && char <= '9' ||
+		char == '_' || char == '-' || char == '.'
+}
+
+func validateHealthCheckAddress(address string) error {
+	if address == "" {
+		return nil
+	}
+
+	parsedURL, err := url.ParseRequestURI(address)
+	if err != nil {
+		return fmt.Errorf("control_plane_health_check_address %q is not a valid URL: %w", address, err)
+	}
+
+	scheme := strings.ToLower(parsedURL.Scheme)
+	if scheme != "http" && scheme != "https" {
+		return fmt.Errorf("control_plane_health_check_address %q has unsupported scheme %q, expected http or https", address, parsedURL.Scheme)
+	}
+
+	return nil
+}

pkg/kubevip/config_validation_test.go

@@ -0,0 +1,70 @@
+package kubevip
+
+import (
+	"strings"
+	"testing"
+)
+
+func TestValidate_HealthCheckAddress(t *testing.T) {
+	tests := []struct {
+		name    string
+		address string
+		wantErr bool
+	}{
+		{"empty address (disabled)", "", false},
+		{"valid http URL", "http://localhost:6443/livez", false},
+		{"valid https URL", "https://localhost:6443/livez", false},
+		{"https with path", "https://127.0.0.1:6443/readyz?verbose", false},
+		{"invalid URL", "not-a-url", true},
+		{"ftp scheme", "ftp://localhost/file", true},
+		{"tcp scheme", "tcp://localhost:6443", true},
+		{"missing scheme", "localhost:6443/livez", true},
+	}
+
+	for _, tt := range tests {
+		t.Run(tt.name, func(t *testing.T) {
+			c := &Config{ControlPlaneHealthCheck: HealthCheck{Address: tt.address}}
+			err := c.Validate()
+			if (err != nil) != tt.wantErr {
+				t.Errorf("Validate() error = %v, wantErr %v", err, tt.wantErr)
+			}
+		})
+	}
+}
+
+func TestValidate_InstanceName(t *testing.T) {
+	tests := []struct {
+		name         string
+		instanceName string
+		wantErr      bool
+	}{
+		{name: "empty uses legacy default", instanceName: "", wantErr: false},
+		{name: "letters digits and separators", instanceName: "release_01.prod-a", wantErr: false},
+		{name: "exact maximum length", instanceName: strings.Repeat("a", instanceNameMaxLength), wantErr: false},
+		{name: "exceeds maximum length", instanceName: strings.Repeat("a", instanceNameMaxLength+1), wantErr: true},
+		{name: "space", instanceName: "release a", wantErr: true},
+		{name: "slash", instanceName: "namespace/release", wantErr: true},
+		{name: "dollar sign", instanceName: "release$a", wantErr: true},
+		{name: "at sign", instanceName: "release@a", wantErr: true},
+		{name: "newline", instanceName: "release\na", wantErr: true},
+		{name: "null byte", instanceName: "release\x00a", wantErr: true},
+		{name: "unicode", instanceName: "rilascio-à", wantErr: true},
+	}
+
+	for _, tt := range tests {
+		t.Run(tt.name, func(t *testing.T) {
+			config := &Config{InstanceName: tt.instanceName}
+			err := config.Validate()
+			if (err != nil) != tt.wantErr {
+				t.Fatalf("Validate() error = %v, wantErr %t", err, tt.wantErr)
+			}
+		})
+	}
+}
+
+func TestInstanceNameLimitReservesNftablesPrefixAndFamilySuffix(t *testing.T) {
+	name := strings.Repeat("a", instanceNameMaxLength)
+	if got := len(egressNftablesTablePrefix + name + egressNftablesTableSuffix); got != nftablesNameMaxLength {
+		t.Fatalf("family-specific table name length = %d, want %d", got, nftablesNameMaxLength)
+	}
+}

pkg/kubevip/constants.go

@@ -0,0 +1,7 @@
+package kubevip
+
+const (
+	LBClassName = "kube-vip.io/kube-vip-class"
+
+	DefaultDHCPBackoffAttempts = 3
+)

pkg/kubevip/labels.go

@@ -0,0 +1,22 @@
+package kubevip
+
+import (
+	"slices"
+)
+
+const (
+	// ServiceProvided is the name of the label that will be added to the node
+	ServiceProvided = "service-provided.kube-vip.io"
+
+	// label used on nodes, which announce the LoadBalancer IP
+	HasIP = "kube-vip.io/has-ip"
+)
+
+var kubevipLabelKeys = []string{
+	ServiceProvided,
+	HasIP,
+}
+
+func GetKeysForCleanup() []string {
+	return slices.Clone(kubevipLabelKeys)
+}

pkg/lease/lease.go

@@ -0,0 +1,270 @@
+package lease
+
+import (
+	"context"
+	"fmt"
+	log "log/slog"
+	"os"
+	"strings"
+	"sync"
+	"sync/atomic"
+
+	"github.com/kube-vip/kube-vip/pkg/kubevip"
+	v1 "k8s.io/api/core/v1"
+)
+
+// Manager is used to manage leases.
+type Manager struct {
+	leases map[string]*Lease
+	lock   sync.Mutex
+}
+
+// NewManager creates new lease manager.
+func NewManager() *Manager {
+	return &Manager{
+		leases: make(map[string]*Lease),
+	}
+}
+
+// Add adds lease to the manager.
+// It returns three values:
+// - lease for the object
+// - isNewObject, which reports if it is a new object that is being handled
+// - isSharedLease, which is true if object shares the lease with another object
+// If object is new but not shared, we should start leaderelection and sync it
+// If object is new and shared, we should only sync it as the leaderelection should be already handled
+// If object is not new we should do nothing
+func (m *Manager) Add(ctx context.Context, id ID) *Lease {
+	m.lock.Lock()
+	defer m.lock.Unlock()
+
+	// A lease whose context is already cancelled cannot be handed out again:
+	// anything derived from it would be cancelled straight away. Replace it.
+	if l, exists := m.leases[id.NamespacedName()]; !exists || l.Ctx.Err() != nil {
+		leaseCtx, leaseCancel := context.WithCancel(ctx)
+		m.leases[id.NamespacedName()] = newLease(leaseCtx, leaseCancel)
+	}
+
+	return m.leases[id.NamespacedName()]
+}
+
+// Delete removes the object from the lease it was added to and cancels that lease
+// once its last object is gone. With a common lease, the siblings that still use
+// it keep it alive.
+//
+// The lease the caller was given has to be passed in, because cleanup is usually
+// deferred to a goroutine that runs long after the object went away. By then the
+// lease of that name may already have been replaced, for instance because the
+// service was torn down and rebuilt, and cancelling the replacement would leave
+// the service unhandled. A stale caller is therefore ignored.
+//
+// Teardown paths have to call this synchronously rather than leaving it to the
+// deferred cleanup: until the lease is out of the map, Add hands the same
+// instance back, so a service that is rebuilt straight away gets parented to a
+// lease that the pending cleanup is about to cancel.
+func (m *Manager) Delete(id ID, objectName string, l *Lease) {
+	m.lock.Lock()
+	defer m.lock.Unlock()
+
+	current := m.currentFor(id, l)
+	if current == nil {
+		return
+	}
+
+	current.delete(objectName)
+	if current.cnt.Load() < 1 {
+		m.retire(id, current)
+	}
+}
+
+// currentFor returns the registered lease for id, or nil when the caller is
+// stale, meaning the lease it holds is no longer the registered one. Callers have
+// to hold m.lock.
+func (m *Manager) currentFor(id ID, l *Lease) *Lease {
+	current, exist := m.leases[id.NamespacedName()]
+	if !exist || (l != nil && current != l) {
+		return nil
+	}
+	return current
+}
+
+// retire cancels the lease and drops it from the manager. Callers have to hold
+// m.lock.
+func (m *Manager) retire(id ID, l *Lease) {
+	l.Cancel()
+	delete(m.leases, id.NamespacedName())
+}
+
+// Get returns lease for the service.
+func (m *Manager) Get(id ID) *Lease {
+	m.lock.Lock()
+	defer m.lock.Unlock()
+
+	if lease, exist := m.leases[id.NamespacedName()]; exist {
+		return lease
+	}
+	return nil
+}
+
+// Lease holds lease data.
+type Lease struct {
+	Ctx      context.Context
+	Cancel   context.CancelFunc
+	Started  chan any
+	services sync.Map
+	cnt      atomic.Int64
+	Elected  atomic.Bool
+	Mtx      sync.Mutex
+	locked   bool
+}
+
+func newLease(ctx context.Context, cancel context.CancelFunc) *Lease {
+	return &Lease{
+		Ctx:     ctx,
+		Cancel:  cancel,
+		Started: make(chan any),
+	}
+}
+
+// Add adds the object to the lease and increments counter
+// it will return true if object was added
+func (l *Lease) Add(name string) bool {
+	if _, exists := l.services.LoadOrStore(name, true); !exists {
+		l.cnt.Add(1)
+		return true
+	}
+	return false
+}
+
+// delete removes the service from the lease and decrements the counter
+func (l *Lease) delete(service string) {
+	if _, exists := l.services.Load(service); exists {
+		l.services.Delete(service)
+		l.cnt.Add(-1)
+	}
+}
+
+func (l *Lease) Lock() {
+	l.Mtx.Lock()
+	l.locked = true
+}
+
+func (l *Lease) Unlock() {
+	if l.locked {
+		l.locked = false
+		l.Mtx.Unlock()
+	}
+}
+
+// ServiceName gets lease name and id for the service.
+func ServiceName(service *v1.Service) (string, string) {
+	name, exists := service.Annotations[kubevip.ServiceLease]
+	if !exists || name == "" {
+		name = fmt.Sprintf("kubevip-%s", service.Name)
+	}
+
+	serviceLeaseParts := strings.Split(name, "/")
+	namespace := service.Namespace
+
+	if len(serviceLeaseParts) > 1 {
+		namespace = serviceLeaseParts[0]
+		name = serviceLeaseParts[1]
+	}
+
+	return namespace, name
+}
+
+func ServiceNamespacedName(service *v1.Service) string {
+	return fmt.Sprintf("%s/%s", service.Namespace, service.Name)
+}
+
+func ObjectName(id ID, suffix string) string {
+	return fmt.Sprintf("%s-%s", id.NamespacedName(), suffix)
+}
+
+func NamespaceName(lease string, c *kubevip.Config) (string, string) {
+	leaseName := lease
+	leasnameParts := strings.Split(lease, "/")
+	var ns string
+	var err error
+	if len(leasnameParts) > 1 {
+		ns = leasnameParts[0]
+		leaseName = leasnameParts[1]
+	} else {
+		ns, err = returnNamespace()
+		if err != nil {
+			log.Warn("unable to auto-detect namespace, dropping to config", "namespace", c.Namespace)
+			ns = c.Namespace
+		}
+	}
+	return ns, leaseName
+}
+
+func returnNamespace() (string, error) {
+	if data, err := os.ReadFile("/var/run/secrets/kubernetes.io/serviceaccount/namespace"); err == nil {
+		if ns := strings.TrimSpace(string(data)); len(ns) > 0 {
+			return ns, nil
+		}
+		return "", err
+	}
+	return "", fmt.Errorf("unable to find Namespace")
+}
+
+type ID interface {
+	Name() string
+	Namespace() string
+	NamespacedName() string
+}
+
+type CommonID struct {
+	namespace string
+	name      string
+}
+
+func NewID(leaseType, namespace, name string) ID {
+	if leaseType == "etcd" {
+		return newEtcdID(namespace, name)
+	}
+	return newKubernetesID(namespace, name)
+}
+
+func newKubernetesID(namespace, name string) ID {
+	return &KubernetesID{
+		CommonID: CommonID{
+			namespace: namespace,
+			name:      name,
+		},
+	}
+}
+func newEtcdID(namespace, name string) ID {
+	return &EtcdID{
+		CommonID: CommonID{
+			namespace: namespace,
+			name:      name,
+		},
+	}
+}
+
+func (c *CommonID) Name() string {
+	return c.name
+}
+
+func (c *CommonID) Namespace() string {
+	return c.namespace
+}
+
+type KubernetesID struct {
+	CommonID
+}
+
+func (k *KubernetesID) NamespacedName() string {
+	return fmt.Sprintf("%s/%s", k.namespace, k.name)
+}
+
+type EtcdID struct {
+	CommonID
+}
+
+func (e *EtcdID) NamespacedName() string {
+	return fmt.Sprintf("%s-%s", e.namespace, e.name)
+}

pkg/loadbalancer/ipvs.go

@@ -16,6 +16,7 @@ import (
 	"github.com/cloudflare/ipvs/netmask"
 	"github.com/kube-vip/kube-vip/pkg/backend"
 	"github.com/kube-vip/kube-vip/pkg/sysctl"
+	"github.com/kube-vip/kube-vip/pkg/utils"
 	"github.com/kube-vip/kube-vip/pkg/vip"
 	"github.com/vishvananda/netlink"
 )
@@ -50,49 +51,61 @@ type IPVSLoadBalancer struct {
 	backendMap          backend.Map
 	interval            int
 	lock                sync.Mutex
-	stop                chan struct{}
-	networkInterface    string
-	leaderCancel        context.CancelFunc
-	signal              chan os.Signal
+	killFunc            func()
 	address             string
 	family              ipvs.AddressFamily
 }
 
-func NewIPVSLB(address string, port uint16, forwardingMethod string, backendHealthCheckInterval int, networkInterface string, leaderCancel context.CancelFunc, signal chan os.Signal) (*IPVSLoadBalancer, error) {
-	log.Info("Starting IPVS LoadBalancer", "address", address)
+func NewIPVSLB(ctx context.Context, network vip.Network, port uint16, forwardingMethod string, backendHealthCheckInterval int,
+	nftables bool, killFunc func(), wg *sync.WaitGroup) (*IPVSLoadBalancer, error) {
+	log.Info("Starting IPVS LoadBalancer", "network", network)
+
+	address := network.IP()
 
 	// Create IPVS client
 	c, err := ipvs.New()
 	if err != nil {
 		log.Error("ensure IPVS kernel modules are loaded")
-		log.Error("Error starting IPVS", "err", err)
-		panic("")
+		log.Error("error starting IPVS", "err", err)
+		return nil, fmt.Errorf("starting IPVS: %w", err)
 	}
 	i, err := c.Info()
 	if err != nil {
 		log.Error("ensure IPVS kernel modules are loaded")
-		log.Error("Error retrieving IPVS info", "err", err)
+		log.Error("error retrieving IPVS info", "err", err)
 		if errors.Is(err, os.ErrPermission) {
 			log.Error("no permission to get IPVS info - please ensure that kube-vip is running with proper capabilities/privileged mode")
 		}
-		panic("")
+		return nil, fmt.Errorf("retrieving IPVS: %w", err)
 	}
 	log.Info("IPVS Loadbalancer enabled", "version", fmt.Sprintf("%d.%d.%d", i.Version[0], i.Version[1], i.Version[2]))
 
 	ip, family := ipAndFamily(address)
 
 	if strings.ToLower(forwardingMethod) == "masquerade" {
-		enableProcSys("/proc/sys/net/ipv4/vs/conntrack", "net.ipv4.vs.conntrack")
+		if err := enableProcSys("/proc/sys/net/ipv4/vs/conntrack", "net.ipv4.vs.conntrack"); err != nil {
+			return nil, err
+		}
 		if family == ipvs.INET6 {
-			enableProcSys("/proc/sys/net/ipv6/conf/all/forwarding", "net.ipv6.conf.all.forwarding")
+			if err := enableProcSys("/proc/sys/net/ipv6/conf/all/forwarding", "net.ipv6.conf.all.forwarding"); err != nil {
+				return nil, err
+			}
 		} else {
-			enableProcSys("/proc/sys/net/ipv4/ip_forward", "net.ipv4.ip_forward")
+			if err := enableProcSys("/proc/sys/net/ipv4/ip_forward", "net.ipv4.ip_forward"); err != nil {
+				return nil, err
+			}
 		}
 	}
 
-	netMask := netmask.MaskFrom(31, 32) // For ipv4
+	netMask := netmask.MaskFrom(31, vip.DefaultMaskIPv4) // For ipv4
 	if family == ipvs.INET6 {
-		netMask = netmask.MaskFrom(128, 128) // For ipv6
+		netMask = netmask.MaskFrom(128, vip.DefaultMaskIPv6) // For ipv6
+	}
+
+	var fwmark uint32
+
+	if nftables && forwardingMethod == "masquerade" && family == ipvs.INET {
+		fwmark = network.IPVSMark()
 	}
 
 	// Generate out API Server LoadBalancer instance
@@ -103,6 +116,7 @@ func NewIPVSLB(address string, port uint16, forwardingMethod string, backendHeal
 		Port:      port,
 		Address:   ip,
 		Scheduler: ROUNDROBIN,
+		FWMark:    fwmark,
 	}
 
 	var m ipvs.ForwardType
@@ -133,34 +147,32 @@ func NewIPVSLB(address string, port uint16, forwardingMethod string, backendHeal
 		forwardingMethod:    m,
 		interval:            backendHealthCheckInterval,
 		backendMap:          make(backend.Map),
-		stop:                make(chan struct{}),
-		networkInterface:    networkInterface,
-		leaderCancel:        leaderCancel,
-		signal:              signal,
+		killFunc:            killFunc,
 		address:             address,
 		family:              family,
 	}
 
-	go lb.healthCheck()
+	wg.Go(func() {
+		lb.healthCheck(ctx)
+	})
 
 	// Return our created load-balancer
 	return lb, nil
 }
 
-func enableProcSys(path, name string) {
+func enableProcSys(path, name string) error {
 	isSet, err := sysctl.EnableProcSys(path)
 	if err != nil {
-		log.Error(fmt.Sprintf("ensuring %s enabled", name), "err", err)
-		panic("")
+		return fmt.Errorf("ensuring %s enabled: %w", name, err)
 	}
 	if isSet {
 		log.Info(fmt.Sprintf("sysctl set %s to 1", name))
 	}
+	return nil
 }
 
 func (lb *IPVSLoadBalancer) RemoveIPVSLB() error {
 	log.Info("Stopping IPVS LoadBalancer", "address", lb.address)
-	close(lb.stop)
 	err := lb.client.RemoveService(lb.loadBalancerService)
 	if err != nil {
 		return fmt.Errorf("error removing existing IPVS service: %v", err)
@@ -183,6 +195,7 @@ func (lb *IPVSLoadBalancer) AddBackend(address string, port uint16) error {
 		if err != nil {
 			log.Error("checking if backend is local", "err", err)
 		}
+		log.Info("checked if backend is local", "addr", address, "local", isLocal)
 	}
 
 	backend := backend.Entry{Addr: address, Port: port, IsLocal: isLocal}
@@ -227,7 +240,7 @@ func (lb *IPVSLoadBalancer) addBackend(address string, port uint16) error {
 			// Fatal error at this point as IPVS is probably not working
 			log.Error("Unable to create an IPVS service, ensure IPVS kernel modules are loaded")
 			log.Error("IPVS service", "err", err)
-			panic("")
+			return utils.WrapPanicError(err, "unable to create an IPVS service")
 
 		}
 		log.Info("load-Balancer services created", "address", lb.addrString(), "port", lb.Port)
@@ -312,8 +325,8 @@ func ipAndFamily(address string) (netip.Addr, ipvs.AddressFamily) {
 	return netip.AddrFrom4([4]byte(ipAddr.To4())), ipvs.INET
 }
 
-func (lb *IPVSLoadBalancer) healthCheck() {
-	backend.Watch(func() {
+func (lb *IPVSLoadBalancer) healthCheck(ctx context.Context) {
+	backend.Watch(ctx, lb.interval, func() {
 		lb.lock.Lock()
 		defer lb.lock.Unlock()
 		for backend, oldStatus := range lb.backendMap {
@@ -338,43 +351,50 @@ func (lb *IPVSLoadBalancer) healthCheck() {
 					lb.backendMap[backend] = newStatus
 				}
 				if lb.forwardingMethod == ipvs.Local && !lb.localBackendExists() {
-					if lb.signal != nil {
-						close(lb.signal)
-					}
-
-					if lb.leaderCancel != nil {
-						lb.leaderCancel()
+					if lb.killFunc != nil {
+						log.Error("no local backends available, restarting kube-vip")
+						lb.killFunc()
 					}
 				}
 			}
 		}
-	}, lb.interval, lb.stop)
+	})
 }
 
 func (lb *IPVSLoadBalancer) isLocal(address string) (bool, error) {
-	link, err := netlink.LinkByName(lb.networkInterface)
-	if err != nil {
-		return false, fmt.Errorf("getting link '%s': %w", lb.networkInterface, err)
-	}
-
-	family := netlink.FAMILY_V6
-	if vip.IsIPv4(address) {
-		family = netlink.FAMILY_V4
-	}
-
 	target := net.ParseIP(address)
 	if target == nil {
-		return false, fmt.Errorf("address '%s' is not a valid IP address", address)
+		return false, fmt.Errorf("unable to parse IP address %s", address)
 	}
 
-	addrs, err := netlink.AddrList(link, family)
+	links, err := netlink.LinkList()
 	if err != nil {
-		return false, fmt.Errorf("listing addresses for link '%s': %w", lb.networkInterface, err)
+		return false, fmt.Errorf("listing links: %w", err)
 	}
 
-	for _, addr := range addrs {
-		if addr.IP.Equal(target) {
-			return true, nil
+	family := netlink.FAMILY_V4
+	if utils.IsIPv6(address) {
+		family = netlink.FAMILY_V6
+	}
+
+	for _, link := range links {
+		if link.Type() == "veth" {
+			continue
+		}
+
+		addrs, err := netlink.AddrList(link, family)
+		if err != nil {
+			log.Error("listing addresses", "link", link.Attrs().Name, "error", err.Error())
+			continue
+		}
+
+		for _, addr := range addrs {
+			if addr.Scope != int(netlink.SCOPE_UNIVERSE) {
+				continue
+			}
+			if addr.IP.Equal(target) {
+				return true, nil
+			}
 		}
 	}
 

pkg/loadbalancer/ipvs_test.go

@@ -6,6 +6,7 @@ import (
 	"testing"
 
 	"github.com/cloudflare/ipvs"
+	"github.com/kube-vip/kube-vip/pkg/utils"
 )
 
 func Test_ipAndFamily(t *testing.T) {
@@ -19,7 +20,7 @@ func Test_ipAndFamily(t *testing.T) {
 		want1 ipvs.AddressFamily
 	}{
 		{
-			name: "IPv4",
+			name: utils.IPv4Family,
 			args: args{
 				address: "192.168.0.20",
 			},
@@ -27,7 +28,7 @@ func Test_ipAndFamily(t *testing.T) {
 			want1: ipvs.INET,
 		},
 		{
-			name: "IPv6",
+			name: utils.IPv6Family,
 			args: args{
 				address: "ff02::3",
 			},

pkg/manager/cluster.go

@@ -1,30 +0,0 @@
-package manager
-
-import (
-	"github.com/pkg/errors"
-
-	"github.com/kube-vip/kube-vip/pkg/cluster"
-	"github.com/kube-vip/kube-vip/pkg/etcd"
-)
-
-func initClusterManager(sm *Manager) (*cluster.Manager, error) {
-	m := &cluster.Manager{
-		SignalChan: sm.signalChan,
-	}
-
-	switch sm.config.LeaderElectionType {
-	case "kubernetes", "":
-		m.KubernetesClient = sm.clientSet
-		m.RetryWatcherClient = sm.rwClientSet
-	case "etcd":
-		client, err := etcd.NewClient(sm.config)
-		if err != nil {
-			return nil, err
-		}
-		m.EtcdClient = client
-	default:
-		return nil, errors.Errorf("invalid LeaderElectionMode %s not supported", sm.config.LeaderElectionType)
-	}
-
-	return m, nil
-}

pkg/manager/manager.go

@@ -9,6 +9,7 @@ import (
 	"path/filepath"
 	"strings"
 	"sync"
+	"sync/atomic"
 	"syscall"
 	"time"
 
@@ -16,19 +17,26 @@ import (
 
 	"github.com/kube-vip/kube-vip/pkg/arp"
 	"github.com/kube-vip/kube-vip/pkg/bgp"
+	"github.com/kube-vip/kube-vip/pkg/cluster"
+	"github.com/kube-vip/kube-vip/pkg/election"
+	"github.com/kube-vip/kube-vip/pkg/iptables"
 	"github.com/kube-vip/kube-vip/pkg/k8s"
 	"github.com/kube-vip/kube-vip/pkg/kubevip"
+	"github.com/kube-vip/kube-vip/pkg/lease"
+	"github.com/kube-vip/kube-vip/pkg/manager/worker"
 	"github.com/kube-vip/kube-vip/pkg/networkinterface"
+	"github.com/kube-vip/kube-vip/pkg/nftables"
+	"github.com/kube-vip/kube-vip/pkg/node"
+	"github.com/kube-vip/kube-vip/pkg/route"
 	"github.com/kube-vip/kube-vip/pkg/services"
 	"github.com/kube-vip/kube-vip/pkg/upnp"
 	"github.com/kube-vip/kube-vip/pkg/utils"
+	"github.com/kube-vip/kube-vip/pkg/vip"
 	"github.com/prometheus/client_golang/prometheus"
 	"k8s.io/client-go/kubernetes"
 	"k8s.io/client-go/rest"
 )
 
-const plunderLock = "plndr-svcs-lock"
-
 // Manager degines the manager of the load-balancing services
 type Manager struct {
 	clientSet   *kubernetes.Clientset
@@ -45,8 +53,7 @@ type Manager struct {
 	// This channel is used to catch an OS signal and trigger a shutdown
 	signalChan chan os.Signal
 
-	// This channel is used to signal a shutdown
-	shutdownChan chan struct{}
+	sigint sync.Once
 
 	svcProcessor *services.Processor
 
@@ -54,10 +61,6 @@ type Manager struct {
 	// from the service watcher
 	countServiceWatchEvent *prometheus.CounterVec
 
-	// This is a prometheus gauge indicating the state of the sessions.
-	// 1 means "ESTABLISHED", 0 means "NOT ESTABLISHED"
-	bgpSessionInfoGauge *prometheus.GaugeVec
-
 	// This mutex is to protect calls from various goroutines
 	mutex sync.Mutex
 
@@ -66,10 +69,27 @@ type Manager struct {
 
 	// This tracks VIPs and performs ARP/NDP advertisement.
 	arpMgr *arp.Manager
+
+	// This tracks node labels and performs label management
+	// implementation will be decided in constructor
+	// based on config.EnableNodeLabeling
+	nodeLabelManager node.LabelManager
+
+	// This variable reports if manager is being closed
+	closing atomic.Bool
+
+	// Will be used for leaderelection when required
+	electionMgr *election.Manager
+
+	// Will handle leases
+	leaseMgr *lease.Manager
+
+	// Will handle routes
+	routeMgr *route.Manager
 }
 
 // New will create a new managing object
-func New(configMap string, config *kubevip.Config) (*Manager, error) {
+func New(ctx context.Context, configMap string, config *kubevip.Config) (*Manager, error) {
 
 	// Instance identity should be the same as k8s node name to ensure better compatibility.
 	// By default k8s sets node name to `hostname -s`,
@@ -84,6 +104,7 @@ func New(configMap string, config *kubevip.Config) (*Manager, error) {
 		}
 		config.NodeName = hostname
 	}
+	config.NodeName = normalizeNodeName(config.NodeName)
 	log.Info("using node name", "name", config.NodeName)
 
 	adminConfigPath := "/etc/kubernetes/admin.conf"
@@ -96,6 +117,21 @@ func New(configMap string, config *kubevip.Config) (*Manager, error) {
 	switch {
 	case config.LeaderElectionType == "etcd":
 		// Do nothing, we don't construct a k8s client for etcd leader election
+	case config.K8sConfigFile != "" && config.K8sConfigFile != adminConfigPath &&
+		config.K8sConfigFile != homeConfigPath && utils.FileExists(config.K8sConfigFile):
+		// An explicitly configured kubeconfig (k8s_config_file env or
+		// --k8sConfigPath) takes precedence over the well-known host paths.
+		// KubernetesAddr, when set, overrides the API endpoint - static pods
+		// on control plane hosts use it to reach their local API server
+		// instead of a VIP that may not be up yet.
+		clientConfig, err = k8s.NewRestConfig(config.K8sConfigFile, false, config.KubernetesAddr)
+		if err != nil {
+			return nil, fmt.Errorf("could not create k8s REST config from file %q: %w", config.K8sConfigFile, err)
+		}
+		if clientset, err = k8s.NewClientset(clientConfig); err != nil {
+			return nil, fmt.Errorf("could not create k8s clientset: %w", err)
+		}
+		log.Info("Using Kubernetes configuration from explicit file", "path", config.K8sConfigFile, "address", config.KubernetesAddr)
 	case utils.FileExists(adminConfigPath):
 		if config.KubernetesAddr != "" {
 			log.Info("k8s address", "address", config.KubernetesAddr)
@@ -181,21 +217,46 @@ func New(configMap string, config *kubevip.Config) (*Manager, error) {
 	// Add Notification for SIGTERM (sent from Kubernetes)
 	signal.Notify(signalChan, syscall.SIGTERM)
 
-	// All watchers and other goroutines should have an additional goroutine that blocks on this, to shut things down
-	shutdownChan := make(chan struct{})
+	// Add Notification for SIGUSR1 (for configuration dump)
+	signal.Notify(signalChan, syscall.SIGUSR1)
 
 	intfMgr := networkinterface.NewManager()
 	arpMgr := arp.NewManager(config)
 
+	// create the node label manager
+	// constructor will decide if it should be a noop or not
+	nodeLabelManager := node.NewManager(config, clientset)
+
 	var bgpServer *bgp.Server
+	// If BGP is enabled then we start a server instance that will broadcast VIPs
 	if config.EnableBGP {
-		bgpServer, err = bgp.NewBGPServer(&config.BGPConfig)
+		var err error
+		// If Annotations have been set then we will look them up
+		if config.Annotations != "" {
+			err = annotationsWatcher(ctx, clientset, rwClientSet, config)
+		} else {
+			log.Debug("No Node annotations to parse")
+		}
+
+		if err != nil {
+			return nil, err
+		}
+		bgpServer, err = bgp.NewBGPServer(config.BGPConfig, log.Level(config.Logging))
 		if err != nil {
 			return nil, fmt.Errorf("creating BGP server: %w", err)
 		}
 	}
 
-	svcProcessor := services.NewServicesProcessor(config, bgpServer, clientset, rwClientSet, shutdownChan, intfMgr, arpMgr)
+	electionMgr, err := election.NewManager(config, clientset, rwClientSet)
+	if err != nil {
+		return nil, fmt.Errorf("creating election manager: %w", err)
+	}
+
+	leaseMgr := lease.NewManager()
+	routeMgr := route.NewManager()
+
+	svcProcessor := services.NewServicesProcessor(config, bgpServer, clientset, rwClientSet,
+		intfMgr, arpMgr, nodeLabelManager, electionMgr, leaseMgr, routeMgr)
 
 	return &Manager{
 		clientSet:   clientset,
@@ -208,35 +269,22 @@ func New(configMap string, config *kubevip.Config) (*Manager, error) {
 			Name:      "all_services_events",
 			Help:      "Count all events fired by the service watcher categorised by event type",
 		}, []string{"type"}),
-		bgpSessionInfoGauge: prometheus.NewGaugeVec(prometheus.GaugeOpts{
-			Namespace: "kube_vip",
-			Subsystem: "manager",
-			Name:      "bgp_session_info",
-			Help:      "Display state of session by setting metric for label value with current state to 1",
-		}, []string{"state", "peer"}),
-		signalChan:   signalChan,
-		shutdownChan: shutdownChan,
-		svcProcessor: svcProcessor,
-		intfMgr:      intfMgr,
-		arpMgr:       arpMgr,
-		bgpServer:    bgpServer,
+		signalChan:       signalChan,
+		svcProcessor:     svcProcessor,
+		intfMgr:          intfMgr,
+		arpMgr:           arpMgr,
+		bgpServer:        bgpServer,
+		nodeLabelManager: nodeLabelManager,
+		electionMgr:      electionMgr,
+		leaseMgr:         leaseMgr,
+		routeMgr:         routeMgr,
 	}, nil
 }
 
 // Start will begin the Manager, which will start services and watch the configmap
-func (sm *Manager) Start() error {
-	// listen for interrupts or the Linux SIGTERM signal and cancel
-	// our context, which the leader election code will observe and
-	// step down
-	sm.signalChan = make(chan os.Signal, 1)
-	// Add Notification for Userland interrupt
-	signal.Notify(sm.signalChan, syscall.SIGINT)
-
-	// Add Notification for SIGTERM (sent from Kubernetes)
-	signal.Notify(sm.signalChan, syscall.SIGTERM)
-
-	// All watchers and other goroutines should have an additional goroutine that blocks on this, to shut things down
-	sm.shutdownChan = make(chan struct{})
+func (sm *Manager) Start(ctx context.Context) error {
+	wg := sync.WaitGroup{}
+	defer wg.Wait()
 
 	// HealthCheck
 	if sm.config.HealthCheckPort != 0 {
@@ -246,7 +294,7 @@ func (sm *Manager) Start() error {
 		http.HandleFunc("/healthz", func(w http.ResponseWriter, _ *http.Request) {
 			fmt.Fprintf(w, "OK")
 		})
-		go func() {
+		wg.Go(func() {
 			server := &http.Server{
 				Addr:              fmt.Sprintf(":%d", sm.config.HealthCheckPort),
 				ReadHeaderTimeout: 3 * time.Second,
@@ -255,25 +303,19 @@ func (sm *Manager) Start() error {
 			if err != nil {
 				log.Error("healthcheck", "unable to start", err)
 			}
-		}()
+		})
 	}
 
-	// If BGP is enabled then we start a server instance that will broadcast VIPs
-	if sm.config.EnableBGP {
-
-		// If Annotations have been set then we will look them up
-		err := sm.parseAnnotations()
-		if err != nil {
-			return err
+	// on exit, clean up the node labels
+	defer func() {
+		if err := sm.nodeLabelManager.CleanUpLabels(10 * time.Second); err != nil {
+			log.Error("CleanUpNodeLabels", "unable to cleanup node labels", err)
 		}
-
-		log.Info("Starting Kube-vip Manager with the BGP engine")
-		return sm.startBGP()
-	}
+	}()
 
 	if sm.config.EnableARP || sm.config.EnableWireguard {
 		if sm.config.EnableUPNP {
-			clients := upnp.GetConnectionClients(context.TODO())
+			clients := upnp.GetConnectionClients(ctx)
 			if len(clients) == 0 {
 				log.Error("Error Enabling UPNP. No Clients found")
 				// Set the struct to false so nothing should use it in future
@@ -287,50 +329,132 @@ func (sm *Manager) Start() error {
 					log.Info("Found UPNP IGD2 Gateway address", "ip", ip)
 				}
 			}
+			// TODO: It would be nice to run the UPNP refresh only on the leader.
+			wg.Go(func() {
+				sm.svcProcessor.RefreshUPNPForwards(ctx)
+			})
 		}
-		// TODO: It would be nice to run the UPNP refresh only on the leader.
-		go sm.svcProcessor.RefreshUPNPForwards()
 	}
 
-	// If ARP is enabled then we start a LeaderElection that will use ARP to advertise VIPs
-	if sm.config.EnableARP {
-		log.Info("Starting Kube-vip Manager with the ARP engine")
-		return sm.startARP(sm.config.NodeName)
+	return sm.startMode(ctx)
+}
+
+// Start will begin the Manager, which will start services and watch the configmap
+func (sm *Manager) startMode(ctx context.Context) error {
+	var cpCluster *cluster.Cluster
+	var err error
+
+	w := worker.New(sm.arpMgr, sm.intfMgr, sm.config, &sm.closing, sm.Kill,
+		sm.svcProcessor, &sm.mutex, sm.clientSet, sm.bgpServer, sm.electionMgr,
+		sm.leaseMgr, sm.routeMgr, sm.nodeLabelManager)
+
+	// use a Go context so we can tell the leaderelection code when we
+	// want to step down
+	wg := sync.WaitGroup{}
+	modeCtx, cancel := context.WithCancel(ctx)
+	defer func() {
+
+		wg.Wait()
+		w.Cleanup()
+		cancel()
+		log.Info("Shutting down Kube-Vip")
+	}()
+
+	log.Info("starting Kube-vip Manager", "mode", w.Name())
+	if err := w.Configure(modeCtx, &wg); err != nil {
+		defer cancel()
+		return fmt.Errorf("failed to configure %s mode: %w", w.Name(), err)
 	}
 
-	if sm.config.EnableWireguard {
-		log.Info("Starting Kube-vip Manager with the Wireguard engine")
-		return sm.startWireguard(sm.config.NodeName)
+	if sm.config.EnableControlPlane {
+		err = w.InitControlPlane()
+		if err != nil {
+			defer cancel()
+			return err
+		}
 	}
 
-	if sm.config.EnableRoutingTable {
-		log.Info("Starting Kube-vip Manager with the Routing Table engine")
-		return sm.startTableMode(sm.config.NodeName)
+	// Shutdown function that will wait on this signal, unless we call it ourselves
+	wg.Go(func() {
+		sm.waitForShutdown(modeCtx, cancel, cpCluster)
+	})
+
+	if sm.config.EnableControlPlane {
+		wg.Go(func() {
+			w.StartControlPlane(modeCtx, sm.electionMgr)
+		})
+	}
+
+	if sm.config.EnableServices {
+		// This will tidy any dangling kube-vip iptables rules
+		if sm.config.EgressClean {
+			tableName := nftables.EgressTableBaseNameForInstance(sm.config.InstanceName)
+			err := nftables.ClearTablesWithName(tableName)
+			if err != nil {
+				log.Warn("[egress]", "mode", "nftables-internal", "clearing error", err)
+			} else {
+				log.Info("[egress]", "mode", "nftables-internal", "tables", "cleared")
+			}
+			// TODO: Deprecate the iptables code v1.2.x
+			err = vip.ClearIPTables(sm.config.EgressWithNftables, sm.config.ServiceNamespace, iptables.ProtocolIPv4)
+			if err != nil {
+				log.Info("[egress]", "legacy-iptables", sm.config.EgressWithNftables, "mode", "IPv4", "error", err)
+			}
+			err = vip.ClearIPTables(sm.config.EgressWithNftables, sm.config.ServiceNamespace, iptables.ProtocolIPv6)
+			if err != nil {
+				log.Info("[egress]", "legacy-iptables", sm.config.EgressWithNftables, "mode", "IPv6", "error", err)
+			}
+		}
+		w.ConfigureServices()
+
+		for {
+			select {
+			case <-modeCtx.Done():
+				return nil
+			default:
+				if err = w.StartServices(modeCtx); err != nil {
+					if utils.IsPanicError(err) {
+						sm.Kill()
+						return fmt.Errorf("failed to reconcile services, non-recoverable error: %w", err)
+					} else {
+						log.Error("failed to reconcile services, restarting", "error", err)
+					}
+				}
+			}
+		}
 	}
 
-	log.Error("prematurely exiting Load-balancer as no modes [ARP/BGP/Wireguard] are enabled")
 	return nil
 }
 
-func returnNameSpace() (string, error) {
-	if data, err := os.ReadFile("/var/run/secrets/kubernetes.io/serviceaccount/namespace"); err == nil {
-		if ns := strings.TrimSpace(string(data)); len(ns) > 0 {
-			return ns, nil
+func (sm *Manager) waitForShutdown(ctx context.Context, cancel context.CancelFunc, cpCluster *cluster.Cluster) {
+	for {
+		sig := <-sm.signalChan
+		switch sig {
+		case syscall.SIGUSR1:
+			log.Info("Received SIGUSR1, dumping configuration")
+			sm.dumpConfiguration(ctx)
+		case syscall.SIGINT, syscall.SIGTERM:
+			sm.closing.Store(true)
+			log.Info("Received kube-vip termination, signaling shutdown")
+			if cpCluster != nil {
+				cpCluster.Stop()
+			}
+			// Cancel the context, which will in turn cancel the leadership and all goroutines
+			cancel()
+			return
 		}
-		return "", err
 	}
-	return "", fmt.Errorf("unable to find Namespace")
 }
 
-func (sm *Manager) parseAnnotations() error {
-	if sm.config.Annotations == "" {
-		log.Debug("No Node annotations to parse")
-		return nil
-	}
-
-	err := sm.annotationsWatcher()
-	if err != nil {
-		return err
-	}
-	return nil
+func (sm *Manager) Kill() {
+	sm.sigint.Do(func() {
+		sm.signalChan <- syscall.SIGINT
+	})
+}
+
+// normalizeNodeName ensures the local machine hostname conforms to
+// Kubernetes RFC1123 node naming conventions (lowercase).
+func normalizeNodeName(name string) string {
+	return strings.ToLower(name)
 }

pkg/manager/manager_arp.go

@@ -1,159 +0,0 @@
-package manager
-
-import (
-	"context"
-	"syscall"
-	"time"
-
-	log "log/slog"
-
-	metav1 "k8s.io/apimachinery/pkg/apis/meta/v1"
-	"k8s.io/client-go/tools/leaderelection"
-	"k8s.io/client-go/tools/leaderelection/resourcelock"
-
-	"github.com/kube-vip/kube-vip/pkg/cluster"
-	"github.com/kube-vip/kube-vip/pkg/iptables"
-	"github.com/kube-vip/kube-vip/pkg/vip"
-)
-
-// Start will begin the Manager, which will start services and watch the configmap
-func (sm *Manager) startARP(id string) error {
-	var cpCluster *cluster.Cluster
-	var ns string
-	var err error
-
-	// use a Go context so we can tell the leaderelection code when we
-	// want to step down
-	ctx, cancel := context.WithCancel(context.Background())
-	defer cancel()
-
-	log.Info("Start ARP/NDP advertisement")
-	go sm.arpMgr.StartAdvertisement(ctx)
-
-	// Shutdown function that will wait on this signal, unless we call it ourselves
-	go func() {
-		<-sm.signalChan
-		log.Info("Received kube-vip termination, signaling shutdown")
-		if sm.config.EnableControlPlane {
-			cpCluster.Stop()
-		}
-		// Close all go routines
-		close(sm.shutdownChan)
-		// Cancel the context, which will in turn cancel the leadership
-		cancel()
-	}()
-
-	if sm.config.EnableControlPlane {
-		cpCluster, err = cluster.InitCluster(sm.config, false, sm.intfMgr, sm.arpMgr)
-		if err != nil {
-			return err
-		}
-
-		clusterManager, err := initClusterManager(sm)
-		if err != nil {
-			return err
-		}
-
-		go func() {
-			err := cpCluster.StartCluster(sm.config, clusterManager, nil)
-			if err != nil {
-				log.Error("starting control plane", "err", err)
-				// Trigger the shutdown of this manager instance
-				sm.signalChan <- syscall.SIGINT
-			}
-		}()
-
-		// Check if we're also starting the services, if not we can sit and wait on the closing channel and return here
-		if !sm.config.EnableServices {
-			<-sm.shutdownChan
-			log.Info("Shutting down Kube-Vip")
-
-			return nil
-		}
-
-		ns = sm.config.Namespace
-	} else {
-
-		ns, err = returnNameSpace()
-		if err != nil {
-			log.Warn("unable to auto-detect namespace, dropping to config", "namespace", sm.config.Namespace)
-			ns = sm.config.Namespace
-		}
-	}
-
-	// This will tidy any dangling kube-vip iptables rules
-	if sm.config.EgressClean {
-		vip.ClearIPTables(sm.config.EgressWithNftables, sm.config.ServiceNamespace, iptables.ProtocolIPv4)
-	}
-
-	// Start a services watcher (all kube-vip pods will watch services), upon a new service
-	// a lock based upon that service is created that they will all leaderElection on
-	if sm.config.EnableServicesElection {
-		log.Info("beginning watching services, leaderelection will happen for every service")
-		err = sm.svcProcessor.StartServicesWatchForLeaderElection(ctx)
-		if err != nil {
-			return err
-		}
-	} else {
-
-		log.Info("beginning services leadership", "namespace", ns, "lock name", sm.config.ServicesLeaseName, "id", id)
-		// we use the Lease lock type since edits to Leases are less common
-		// and fewer objects in the cluster watch "all Leases".
-		lock := &resourcelock.LeaseLock{
-			LeaseMeta: metav1.ObjectMeta{
-				Name:      sm.config.ServicesLeaseName,
-				Namespace: ns,
-			},
-			Client: sm.clientSet.CoordinationV1(),
-			LockConfig: resourcelock.ResourceLockConfig{
-				Identity: id,
-			},
-		}
-
-		// start the leader election code loop
-		leaderelection.RunOrDie(ctx, leaderelection.LeaderElectionConfig{
-			Lock: lock,
-			// IMPORTANT: you MUST ensure that any code you have that
-			// is protected by the lease must terminate **before**
-			// you call cancel. Otherwise, you could have a background
-			// loop still running and another process could
-			// get elected before your background loop finished, violating
-			// the stated goal of the lease.
-			ReleaseOnCancel: true,
-			LeaseDuration:   time.Duration(sm.config.LeaseDuration) * time.Second,
-			RenewDeadline:   time.Duration(sm.config.RenewDeadline) * time.Second,
-			RetryPeriod:     time.Duration(sm.config.RetryPeriod) * time.Second,
-			Callbacks: leaderelection.LeaderCallbacks{
-				OnStartedLeading: func(ctx context.Context) {
-					err = sm.svcProcessor.ServicesWatcher(ctx, sm.svcProcessor.SyncServices)
-					if err != nil {
-						log.Error("service watcher", "err", err)
-						panic("") // TODO: - emulating log.fatal here
-					}
-				},
-				OnStoppedLeading: func() {
-					// we can do cleanup here
-					sm.mutex.Lock()
-					defer sm.mutex.Unlock()
-					log.Info("leader lost", "new leader", id)
-					sm.svcProcessor.Stop()
-
-					log.Error("lost leadership, restarting kube-vip")
-					panic("") // TODO: - emulating log.fatal here
-				},
-				OnNewLeader: func(identity string) {
-					// we're notified when new leader elected
-					if sm.config.EnableNodeLabeling {
-						applyNodeLabel(sm.clientSet, sm.config.Address, id, identity)
-					}
-					if identity == id {
-						// I just got the lock
-						return
-					}
-					log.Info("new leader elected", "new leader", identity)
-				},
-			},
-		})
-	}
-	return nil
-}

pkg/manager/manager_bgp.go

@@ -1,116 +0,0 @@
-package manager
-
-import (
-	"context"
-	"fmt"
-	"syscall"
-
-	log "log/slog"
-
-	"github.com/kube-vip/kube-vip/pkg/bgp"
-	"github.com/kube-vip/kube-vip/pkg/cluster"
-	api "github.com/osrg/gobgp/v3/api"
-	"github.com/prometheus/client_golang/prometheus"
-)
-
-// Start will begin the Manager, which will start services and watch the configmap
-func (sm *Manager) startBGP() error {
-	var cpCluster *cluster.Cluster
-	// var ns string
-	var err error
-
-	if sm.bgpServer == nil {
-		sm.bgpServer, err = bgp.NewBGPServer(&sm.config.BGPConfig)
-		if err != nil {
-			return fmt.Errorf("creating BGP server: %w", err)
-		}
-	}
-
-	log.Info("Starting the BGP server to advertise VIP routes to BGP peers")
-	if err := sm.bgpServer.Start(func(p *api.WatchEventResponse_PeerEvent) {
-		ipaddr := p.GetPeer().GetState().GetNeighborAddress()
-		port := uint64(179)
-		peerDescription := fmt.Sprintf("%s:%d", ipaddr, port)
-
-		for stateName, stateValue := range api.PeerState_SessionState_value {
-			metricValue := 0.0
-			if stateValue == int32(p.GetPeer().GetState().GetSessionState().Number()) {
-				metricValue = 1
-			}
-
-			sm.bgpSessionInfoGauge.With(prometheus.Labels{
-				"state": stateName,
-				"peer":  peerDescription,
-			}).Set(metricValue)
-		}
-	}); err != nil {
-		return fmt.Errorf("starting BGP server: %w", err)
-	}
-
-	// use a Go context so we can tell the leaderelection code when we
-	// want to step down
-	ctx, cancel := context.WithCancel(context.Background())
-	defer cancel()
-
-	// Defer a function to check if the bgpServer has been created and if so attempt to close it
-	defer func() {
-		if sm.bgpServer != nil {
-			sm.bgpServer.Close()
-		}
-	}()
-
-	// Shutdown function that will wait on this signal, unless we call it ourselves
-	go func() {
-		<-sm.signalChan
-		log.Info("Received termination, signaling shutdown")
-		if sm.config.EnableControlPlane {
-			if cpCluster != nil {
-				cpCluster.Stop()
-			}
-		}
-		// Cancel the context, which will in turn cancel the leadership
-		cancel()
-	}()
-
-	if sm.config.EnableControlPlane {
-		cpCluster, err = cluster.InitCluster(sm.config, false, sm.intfMgr, sm.arpMgr)
-		if err != nil {
-			return err
-		}
-
-		clusterManager, err := initClusterManager(sm)
-		if err != nil {
-			return err
-		}
-
-		go func() {
-			if sm.config.EnableLeaderElection {
-				err = cpCluster.StartCluster(sm.config, clusterManager, sm.bgpServer)
-			} else {
-				err = cpCluster.StartVipService(sm.config, clusterManager, sm.bgpServer)
-			}
-			if err != nil {
-				log.Error("Control Plane", "err", err)
-				// Trigger the shutdown of this manager instance
-				sm.signalChan <- syscall.SIGINT
-			}
-		}()
-
-		// Check if we're also starting the services, if not we can sit and wait on the closing channel and return here
-		if !sm.config.EnableServices {
-			<-sm.signalChan
-			log.Info("Shutting down Kube-Vip")
-
-			return nil
-		}
-	}
-
-	err = sm.svcProcessor.ServicesWatcher(ctx, sm.svcProcessor.SyncServices)
-	if err != nil {
-		return err
-	}
-
-	log.Info("Shutting down Kube-Vip")
-
-	return nil
-}

pkg/manager/manager_dump.go

@@ -0,0 +1,290 @@
+package manager
+
+import (
+	"context"
+	"fmt"
+	"os"
+	"time"
+
+	"github.com/kube-vip/kube-vip/pkg/nftables"
+	v1 "k8s.io/api/core/v1"
+	metav1 "k8s.io/apimachinery/pkg/apis/meta/v1"
+)
+
+// dumpConfiguration prints the current configuration to stdout when SIGUSR1 is received
+func (sm *Manager) dumpConfiguration(ctx context.Context) {
+	sm.mutex.Lock()
+	defer sm.mutex.Unlock()
+
+	fmt.Printf("\n")
+	fmt.Printf("================================================================================\n")
+	fmt.Printf("                   KUBE-VIP CONFIGURATION DUMP\n")
+	fmt.Printf("================================================================================\n")
+	fmt.Printf("Timestamp: %s\n", time.Now().Format(time.RFC3339))
+	fmt.Printf("Node Name: %s\n", sm.config.NodeName)
+	fmt.Printf("Process ID: %d\n", os.Getpid())
+	fmt.Printf("================================================================================\n")
+	fmt.Printf("\n")
+
+	sm.dumpConfigSection()
+	sm.dumpBGPSection(ctx)
+	sm.dumpARPSection()
+	sm.dumpServicesSection(ctx)
+	sm.dumpNetworkInterfacesSection()
+	sm.dumpLeaderElectionSection()
+	sm.dumpRuntimeSection()
+	sm.dumpNFTablesSection()
+	fmt.Printf("================================================================================\n")
+	fmt.Printf("                   END OF CONFIGURATION DUMP\n")
+	fmt.Printf("================================================================================\n")
+	fmt.Printf("\n")
+}
+
+func (sm *Manager) dumpConfigSection() {
+	fmt.Printf("--- BASIC CONFIGURATION ---\n")
+	fmt.Printf("VIP: %s\n", sm.config.Address)
+	fmt.Printf("VIP Subnet: %s\n", sm.config.VIPSubnet)
+	fmt.Printf("Port: %d\n", sm.config.Port)
+	fmt.Printf("Namespace: %s\n", sm.config.Namespace)
+	fmt.Printf("Instance Name: %s\n", sm.config.InstanceName)
+	fmt.Printf("Service Namespace: %s\n", sm.config.ServiceNamespace)
+	fmt.Printf("Interface: %s\n", sm.config.Interface)
+	fmt.Printf("Services Interface: %s\n", sm.config.ServicesInterface)
+	fmt.Printf("Single Node Mode: %t\n", sm.config.SingleNode)
+	fmt.Printf("Start As Leader: %t\n", sm.config.StartAsLeader)
+	fmt.Printf("\n")
+}
+
+func (sm *Manager) dumpBGPSection(ctx context.Context) {
+	fmt.Printf("--- BGP CONFIGURATION ---\n")
+	fmt.Printf("BGP Enabled: %t\n", sm.config.EnableBGP)
+	if sm.config.EnableBGP {
+		fmt.Printf("BGP AS: %d\n", sm.config.BGPConfig.AS)
+		fmt.Printf("BGP Router ID: %s\n", sm.config.BGPConfig.RouterID)
+		fmt.Printf("BGP Source IP: %s\n", sm.config.BGPConfig.SourceIP)
+		fmt.Printf("BGP Source Interface: %s\n", sm.config.BGPConfig.SourceIF)
+		fmt.Printf("BGP Hold Time: %d\n", sm.config.BGPConfig.HoldTime)
+		fmt.Printf("BGP Keepalive Interval: %d\n", sm.config.BGPConfig.KeepaliveInterval)
+		fmt.Printf("BGP Peers: %d\n", len(sm.config.BGPConfig.Peers))
+		for i, peer := range sm.config.BGPConfig.Peers {
+			fmt.Printf("  Peer %d: %s:%d (AS: %d, MultiHop: %t)\n",
+				i+1, peer.Address, peer.Port, peer.AS, peer.MultiHop)
+		}
+		fmt.Printf("\n--- ACTIVE BGP RIB STATE ---\n")
+		sm.dumpBGPRoutes(ctx)
+	}
+	fmt.Printf("\n")
+}
+
+func (sm *Manager) dumpARPSection() {
+	fmt.Printf("--- ARP/NDP CONFIGURATION ---\n")
+	fmt.Printf("ARP Enabled: %t\n", sm.config.EnableARP)
+	if sm.config.EnableARP {
+		fmt.Printf("ARP Broadcast Rate: %d\n", sm.config.ArpBroadcastRate)
+	}
+	fmt.Printf("Wireguard Enabled: %t\n", sm.config.EnableWireguard)
+	fmt.Printf("Routing Table Enabled: %t\n", sm.config.EnableRoutingTable)
+	if sm.config.EnableRoutingTable {
+		fmt.Printf("Routing Table ID: %d\n", sm.config.RoutingTableID)
+		fmt.Printf("Routing Protocol: %d\n", sm.config.RoutingProtocol)
+		fmt.Printf("Clean Routing Table: %t\n", sm.config.CleanRoutingTable)
+	}
+	fmt.Printf("\n")
+}
+
+func (sm *Manager) dumpServicesSection(ctx context.Context) {
+	fmt.Printf("--- SERVICES CONFIGURATION ---\n")
+	fmt.Printf("Services Enabled: %t\n", sm.config.EnableServices)
+	if sm.config.EnableServices {
+		fmt.Printf("Services Election: %t\n", sm.config.EnableServicesElection)
+		fmt.Printf("Load Balancer Class Only: %t\n", sm.config.LoadBalancerClassOnly)
+		fmt.Printf("Load Balancer Class Name: %s\n", sm.config.LoadBalancerClassName)
+		fmt.Printf("Disable Service Updates: %t\n", sm.config.DisableServiceUpdates)
+		fmt.Printf("Enable Endpoints: %t\n", sm.config.EnableEndpoints)
+		fmt.Printf("Service Security Enabled: %t\n", sm.config.EnableServiceSecurity)
+
+		if sm.svcProcessor != nil {
+			instances := sm.svcProcessor.ServiceInstances
+			fmt.Printf("Kube-vip Active Service Instances: %d\n", len(instances))
+			for i, inst := range instances {
+				if inst.ServiceSnapshot != nil {
+					svc := inst.ServiceSnapshot
+					vipConfigs := ""
+					for j, cfg := range svc.Status.LoadBalancer.Ingress {
+						if j > 0 {
+							vipConfigs += ", "
+						}
+						vipConfigs += cfg.IP
+					}
+					fmt.Printf("  Service %d: %s/%s (Type: %s, VIPs: %s)\n",
+						i+1, svc.Namespace, svc.Name, svc.Spec.Type, vipConfigs)
+				}
+			}
+		}
+		if sm.clientSet != nil {
+			fmt.Println()
+			// Kubernetes configuration
+			fmt.Println("--- KUBERNETES CONFIGURATION (SERVICES/ENDPOINTSLICES) ---")
+
+			fmt.Println("Service Configuration:")
+			svcList, err := sm.clientSet.CoreV1().Services(v1.NamespaceAll).List(ctx, metav1.ListOptions{})
+			if err != nil {
+				fmt.Println("Unable to retrieve all Services")
+			} else {
+				for x := range svcList.Items {
+
+					// Build all addresses
+					vipConfigs := ""
+					for j, cfg := range svcList.Items[x].Status.LoadBalancer.Ingress {
+						if j > 0 {
+							vipConfigs += ", "
+						}
+						vipConfigs += cfg.IP
+					}
+					fmt.Printf("Name=%s, UUID=%s, Addresses=%s\n", svcList.Items[x].Name, string(svcList.Items[x].UID), vipConfigs)
+				}
+				fmt.Println()
+			}
+
+			fmt.Println("EndpointSlice Configuration (note endpoint names have -XXXXX prefixed):")
+			epList, err := sm.clientSet.DiscoveryV1().EndpointSlices(v1.NamespaceAll).List(ctx, metav1.ListOptions{})
+			if err != nil {
+				fmt.Println("Unable to retrieve all EndpointSlices")
+			} else {
+				for x := range epList.Items {
+					// Build all addresses
+					fmt.Printf("  Endpoint Slice Name: %s\n", epList.Items[x].Name)
+					for _, ep := range epList.Items[x].Endpoints {
+						endpoints := ""
+						for i, addresses := range ep.Addresses {
+							if i > 0 {
+								endpoints += ", "
+							}
+							endpoints += addresses
+						}
+						nodeName := "Unknown"
+						targetPod := "Unknown"
+						if ep.NodeName != nil {
+							nodeName = *ep.NodeName
+						}
+						if ep.TargetRef != nil {
+							targetPod = ep.TargetRef.Name
+						}
+						fmt.Printf("\tNode: %s, Target Pod:%s, Addresses: %s\n", nodeName, targetPod, endpoints)
+					}
+				}
+			}
+		}
+	}
+	fmt.Printf("\n")
+}
+
+func (sm *Manager) dumpNetworkInterfacesSection() {
+	fmt.Printf("--- NETWORK INTERFACES ---\n")
+	fmt.Printf("Network Interface Manager: %t\n", sm.intfMgr != nil)
+	fmt.Printf("ARP Manager: %t\n", sm.arpMgr != nil)
+	fmt.Printf("\n")
+}
+
+func (sm *Manager) dumpLeaderElectionSection() {
+	fmt.Printf("--- LEADER ELECTION CONFIGURATION ---\n")
+	fmt.Printf("Control Plane Enabled: %t\n", sm.config.EnableControlPlane)
+	if sm.config.EnableControlPlane {
+		fmt.Printf("Detect Control Plane: %t\n", sm.config.DetectControlPlane)
+	}
+	fmt.Printf("Leader Election Type: %s\n", sm.config.LeaderElectionType)
+	fmt.Printf("Leader Election Enabled: %t\n", sm.config.EnableLeaderElection)
+	if sm.config.EnableLeaderElection {
+		fmt.Printf("Lease Name: %s\n", sm.config.LeaseName)
+		fmt.Printf("Lease Duration: %d seconds\n", sm.config.LeaseDuration)
+		fmt.Printf("Renew Deadline: %d seconds\n", sm.config.RenewDeadline)
+		fmt.Printf("Retry Period: %d seconds\n", sm.config.RetryPeriod)
+	}
+	fmt.Printf("Services Lease Name: %s\n", sm.config.ServicesLeaseName)
+	fmt.Printf("Node Labeling Enabled: %t\n", sm.config.EnableNodeLabeling)
+	fmt.Printf("\n")
+}
+
+func (sm *Manager) dumpRuntimeSection() {
+	fmt.Printf("--- RUNTIME STATISTICS ---\n")
+	fmt.Printf("Load Balancer Enabled: %t\n", sm.config.EnableLoadBalancer)
+	if sm.config.EnableLoadBalancer {
+		fmt.Printf("Load Balancer Port: %d\n", sm.config.LoadBalancerPort)
+		fmt.Printf("Load Balancer Forwarding Method: %s\n", sm.config.LoadBalancerForwardingMethod)
+		fmt.Printf("Load Balancers Configured: %d\n", len(sm.config.LoadBalancers))
+	}
+	fmt.Printf("Prometheus HTTP Server: %s\n", sm.config.PrometheusHTTPServer)
+	fmt.Printf("Health Check Port: %d\n", sm.config.HealthCheckPort)
+	fmt.Printf("UPNP Enabled: %t\n", sm.config.EnableUPNP)
+	fmt.Printf("Egress Clean Enabled: %t\n", sm.config.EgressClean)
+	if sm.config.EgressClean {
+		fmt.Printf("Egress with nftables: %t\n", sm.config.EgressWithNftables)
+		fmt.Printf("Egress nftables table name: %s\n", nftables.EgressTableBaseNameForInstance(sm.config.InstanceName))
+		fmt.Printf("Egress Pod CIDR: %s\n", sm.config.EgressPodCidr)
+		fmt.Printf("Egress Service CIDR: %s\n", sm.config.EgressServiceCidr)
+	}
+	fmt.Printf("\n")
+}
+
+func (sm *Manager) dumpNFTablesSection() {
+	fmt.Printf("--- NFTABLES CONFIGURATION ---\n")
+	chains, err := nftables.ListChains()
+	if err != nil {
+		fmt.Printf("Unable to retrieve NFTables chains, error=%s", err)
+	}
+	for x := range chains {
+		fmt.Printf("Chain: %s\n", chains[x])
+	}
+	fmt.Println()
+}
+
+func (sm *Manager) dumpBGPRoutes(ctx context.Context) {
+	if sm.bgpServer == nil {
+		fmt.Printf("  BGP Server instance is inactive or uninitialized\n")
+		return
+	}
+
+	// Create a short-lived execution window so a stuck BGP loop won't hang the entire SIGUSR1 routine
+	queryCtx, cancel := context.WithTimeout(ctx, 3*time.Second)
+	defer cancel()
+
+	for _, isIPv6 := range []bool{false, true} {
+		label := "IPv4"
+		if isIPv6 {
+			label = "IPv6"
+		}
+
+		routes, err := sm.bgpServer.ListAdvertisedRoutes(queryCtx, isIPv6)
+		if err != nil {
+			fmt.Printf("  Error fetching %s routes: %v\n", label, err)
+			continue
+		}
+
+		if len(routes) == 0 {
+			fmt.Printf("  No %s routes found in global RIB\n", label)
+			continue
+		}
+
+		fmt.Printf("  %-18s | %-15s | %s\n", "Prefix", "Next Hop", "Discovered/Updated")
+		fmt.Printf("  ------------------------------------------------------------\n")
+
+		for _, dest := range routes {
+			for _, path := range dest.Paths {
+				nextHop := "N/A"
+				if path.NeighborIp != "" {
+					nextHop = path.NeighborIp
+				}
+
+				var timeStr string
+				if path.Age != nil {
+					timeStr = path.Age.AsTime().Format("15:04:05")
+				} else {
+					timeStr = "Unknown"
+				}
+
+				fmt.Printf("  %-18s | %-15s | %s\n", dest.Prefix, nextHop, timeStr)
+			}
+		}
+		fmt.Println()
+	}
+}

pkg/manager/manager_dump_test.go

@@ -0,0 +1,242 @@
+package manager
+
+import (
+	"bytes"
+	"context"
+	"io"
+	"os"
+	"testing"
+
+	"github.com/kube-vip/kube-vip/pkg/kubevip"
+	"github.com/stretchr/testify/assert"
+)
+
+func TestDumpConfiguration(t *testing.T) {
+	config := &kubevip.Config{
+		Address:            "192.168.1.100",
+		Interface:          "eth0",
+		Port:               6443,
+		EnableARP:          true,
+		EnableBGP:          false,
+		EnableControlPlane: true,
+		EnableServices:     true,
+		LeaderElectionType: "kubernetes",
+		Namespace:          "kube-system",
+		NodeName:           "test-node",
+		KubernetesLeaderElection: kubevip.KubernetesLeaderElection{
+			EnableLeaderElection: true,
+			LeaseName:            "test-lease",
+		},
+	}
+
+	mgr := &Manager{
+		config: config,
+	}
+
+	old := os.Stdout
+	r, w, _ := os.Pipe()
+	os.Stdout = w
+
+	mgr.dumpConfiguration(context.TODO())
+
+	w.Close()
+	os.Stdout = old
+
+	var buf bytes.Buffer
+	_, err := io.Copy(&buf, r)
+	assert.NoError(t, err, "io.Copy should not return error")
+	output := buf.String()
+
+	assert.Contains(t, output, "KUBE-VIP CONFIGURATION DUMP", "should contain header")
+	assert.Contains(t, output, "Node Name: test-node", "should contain node name")
+	assert.Contains(t, output, "VIP: 192.168.1.100", "should contain VIP address")
+	assert.Contains(t, output, "Interface: eth0", "should contain interface")
+	assert.Contains(t, output, "Port: 6443", "should contain port")
+	assert.Contains(t, output, "ARP Enabled: true", "should contain ARP status")
+	assert.Contains(t, output, "BGP Enabled: false", "should contain BGP status")
+}
+
+func TestDumpConfigSection(t *testing.T) {
+	config := &kubevip.Config{
+		Address:            "192.168.1.100",
+		Interface:          "eth0",
+		Port:               6443,
+		VIPSubnet:          "/24",
+		EnableARP:          true,
+		EnableBGP:          false,
+		EnableControlPlane: true,
+		EnableServices:     true,
+		LeaderElectionType: "kubernetes",
+		Namespace:          "kube-system",
+		KubernetesLeaderElection: kubevip.KubernetesLeaderElection{
+			EnableLeaderElection: true,
+			LeaseName:            "test-lease",
+		},
+	}
+
+	mgr := &Manager{config: config}
+
+	old := os.Stdout
+	r, w, _ := os.Pipe()
+	os.Stdout = w
+
+	mgr.dumpConfigSection()
+
+	w.Close()
+	os.Stdout = old
+
+	var buf bytes.Buffer
+	_, err := io.Copy(&buf, r)
+	assert.NoError(t, err, "io.Copy should not return error")
+	output := buf.String()
+
+	assert.Contains(t, output, "--- BASIC CONFIGURATION ---")
+	assert.Contains(t, output, "VIP: 192.168.1.100")
+	assert.Contains(t, output, "VIP Subnet: /24")
+	assert.Contains(t, output, "Interface: eth0")
+	assert.Contains(t, output, "Port: 6443")
+	assert.Contains(t, output, "Namespace: kube-system")
+	assert.Contains(t, output, "Single Node Mode: false")
+	assert.Contains(t, output, "Start As Leader: false")
+}
+
+func TestDumpBGPSection(t *testing.T) {
+	t.Run("BGP disabled", func(t *testing.T) {
+		config := &kubevip.Config{
+			EnableBGP: false,
+		}
+		mgr := &Manager{config: config}
+
+		old := os.Stdout
+		r, w, _ := os.Pipe()
+		os.Stdout = w
+
+		mgr.dumpBGPSection(t.Context())
+
+		w.Close()
+		os.Stdout = old
+
+		var buf bytes.Buffer
+		_, err := io.Copy(&buf, r)
+		assert.NoError(t, err, "io.Copy should not return error")
+		output := buf.String()
+
+		assert.Contains(t, output, "BGP Enabled: false")
+	})
+
+	t.Run("BGP enabled", func(t *testing.T) {
+		config := &kubevip.Config{
+			EnableBGP: true,
+			BGPConfig: kubevip.BGPConfig{
+				RouterID: "192.168.1.1",
+				AS:       65000,
+				Peers: []kubevip.BGPPeer{
+					{Address: "192.168.1.2", AS: 65001},
+					{Address: "192.168.1.3", AS: 65002},
+				},
+			},
+		}
+		mgr := &Manager{config: config}
+
+		old := os.Stdout
+		r, w, _ := os.Pipe()
+		os.Stdout = w
+
+		mgr.dumpBGPSection(t.Context())
+
+		w.Close()
+		os.Stdout = old
+
+		var buf bytes.Buffer
+		_, err := io.Copy(&buf, r)
+		assert.NoError(t, err, "io.Copy should not return error")
+		output := buf.String()
+
+		assert.Contains(t, output, "BGP Enabled: true")
+		assert.Contains(t, output, "BGP Router ID: 192.168.1.1")
+		assert.Contains(t, output, "BGP AS: 65000")
+		assert.Contains(t, output, "BGP Peers: 2")
+
+		assert.Contains(t, output, "--- ACTIVE BGP RIB STATE ---")
+		assert.Contains(t, output, "BGP Server instance is inactive or uninitialized")
+	})
+}
+
+func TestDumpARPSection(t *testing.T) {
+	t.Run("ARP disabled", func(t *testing.T) {
+		config := &kubevip.Config{
+			EnableARP: false,
+		}
+		mgr := &Manager{config: config}
+
+		old := os.Stdout
+		r, w, _ := os.Pipe()
+		os.Stdout = w
+
+		mgr.dumpARPSection()
+
+		w.Close()
+		os.Stdout = old
+
+		var buf bytes.Buffer
+		_, err := io.Copy(&buf, r)
+		assert.NoError(t, err, "io.Copy should not return error")
+		output := buf.String()
+
+		assert.Contains(t, output, "ARP Enabled: false")
+	})
+
+	t.Run("ARP enabled", func(t *testing.T) {
+		config := &kubevip.Config{
+			EnableARP:        true,
+			ArpBroadcastRate: 5,
+		}
+		mgr := &Manager{config: config}
+
+		old := os.Stdout
+		r, w, _ := os.Pipe()
+		os.Stdout = w
+
+		mgr.dumpARPSection()
+
+		w.Close()
+		os.Stdout = old
+
+		var buf bytes.Buffer
+		_, err := io.Copy(&buf, r)
+		assert.NoError(t, err, "io.Copy should not return error")
+		output := buf.String()
+
+		assert.Contains(t, output, "ARP Enabled: true")
+		assert.Contains(t, output, "ARP Broadcast Rate: 5")
+	})
+}
+
+func TestDumpRuntimeSection(t *testing.T) {
+	config := &kubevip.Config{
+		EnableLoadBalancer:   false,
+		PrometheusHTTPServer: "",
+		HealthCheckPort:      0,
+		EnableUPNP:           false,
+		EgressClean:          false,
+	}
+	mgr := &Manager{config: config}
+
+	old := os.Stdout
+	r, w, _ := os.Pipe()
+	os.Stdout = w
+
+	mgr.dumpRuntimeSection()
+
+	w.Close()
+	os.Stdout = old
+
+	var buf bytes.Buffer
+	_, err := io.Copy(&buf, r)
+	assert.NoError(t, err, "io.Copy should not return error")
+	output := buf.String()
+
+	assert.Contains(t, output, "--- RUNTIME STATISTICS ---", "should contain runtime section header")
+	assert.Contains(t, output, "Load Balancer Enabled: false", "should contain load balancer status")
+	assert.Contains(t, output, "UPNP Enabled: false", "should contain UPNP status")
+}

pkg/manager/manager_table.go

@@ -1,185 +0,0 @@
-package manager
-
-import (
-	"context"
-	"fmt"
-	"syscall"
-	"time"
-
-	log "log/slog"
-
-	"github.com/kube-vip/kube-vip/pkg/cluster"
-	"github.com/kube-vip/kube-vip/pkg/endpoints"
-	"github.com/kube-vip/kube-vip/pkg/iptables"
-	"github.com/kube-vip/kube-vip/pkg/vip"
-	"github.com/vishvananda/netlink"
-	metav1 "k8s.io/apimachinery/pkg/apis/meta/v1"
-	"k8s.io/client-go/tools/leaderelection"
-	"k8s.io/client-go/tools/leaderelection/resourcelock"
-)
-
-// Start will begin the Manager, which will start services and watch the configmap
-func (sm *Manager) startTableMode(id string) error {
-	var cpCluster *cluster.Cluster
-	var err error
-
-	// use a Go context so we can tell the leaderelection code when we
-	// want to step down
-	ctx, cancel := context.WithCancel(context.Background())
-	defer cancel()
-	log.Info("destination for routes", "table", sm.config.RoutingTableID, "protocol", sm.config.RoutingProtocol)
-
-	if sm.config.CleanRoutingTable {
-		go func() {
-			// we assume that after 10s all services should be configured so we can delete redundant routes
-			time.Sleep(time.Second * 10)
-			if err := sm.cleanRoutes(); err != nil {
-				log.Error("error checking for old routes", "err", err)
-			}
-		}()
-	}
-
-	if sm.config.EgressClean {
-		vip.ClearIPTables(sm.config.EgressWithNftables, sm.config.ServiceNamespace, iptables.ProtocolIPv4)
-		vip.ClearIPTables(sm.config.EgressWithNftables, sm.config.ServiceNamespace, iptables.ProtocolIPv6)
-		log.Debug("IPtables rules cleaned on startup")
-	}
-
-	// Shutdown function that will wait on this signal, unless we call it ourselves
-	go func() {
-		<-sm.signalChan
-		log.Info("Received kube-vip termination, signaling shutdown")
-		if sm.config.EnableControlPlane {
-			cpCluster.Stop()
-		}
-
-		// Cancel the context, which will in turn cancel the leadership
-		cancel()
-	}()
-
-	if sm.config.EnableControlPlane {
-		cpCluster, err = cluster.InitCluster(sm.config, false, sm.intfMgr, sm.arpMgr)
-		if err != nil {
-			return fmt.Errorf("cluster initialization error: %w", err)
-		}
-
-		clusterManager, err := initClusterManager(sm)
-		if err != nil {
-			return fmt.Errorf("cluster manager initialization error: %w", err)
-		}
-
-		if err := cpCluster.StartVipService(sm.config, clusterManager, nil); err != nil {
-			log.Error("Control Plane", "err", err)
-			// Trigger the shutdown of this manager instance
-			sm.signalChan <- syscall.SIGINT
-		}
-	} else {
-		ns, err := returnNameSpace()
-		if err != nil {
-			log.Warn("unable to auto-detect namespace", "dropping to", sm.config.Namespace)
-			ns = sm.config.Namespace
-		}
-
-		// Start a services watcher (all kube-vip pods will watch services), upon a new service
-		// a lock based upon that service is created that they will all leaderElection on
-		if sm.config.EnableServicesElection {
-			log.Info("beginning watching services, leaderelection will happen for every service")
-			err = sm.svcProcessor.StartServicesWatchForLeaderElection(ctx)
-			if err != nil {
-				return err
-			}
-		} else if sm.config.EnableLeaderElection {
-
-			log.Info("beginning services leadership", "namespace", ns, "lock name", plunderLock, "id", id)
-			// we use the Lease lock type since edits to Leases are less common
-			// and fewer objects in the cluster watch "all Leases".
-			lock := &resourcelock.LeaseLock{
-				LeaseMeta: metav1.ObjectMeta{
-					Name:      plunderLock,
-					Namespace: ns,
-				},
-				Client: sm.clientSet.CoordinationV1(),
-				LockConfig: resourcelock.ResourceLockConfig{
-					Identity: id,
-				},
-			}
-			// start the leader election code loop
-			leaderelection.RunOrDie(ctx, leaderelection.LeaderElectionConfig{
-				Lock: lock,
-				// IMPORTANT: you MUST ensure that any code you have that
-				// is protected by the lease must terminate **before**
-				// you call cancel. Otherwise, you could have a background
-				// loop still running and another process could
-				// get elected before your background loop finished, violating
-				// the stated goal of the lease.
-				ReleaseOnCancel: true,
-				LeaseDuration:   time.Duration(sm.config.LeaseDuration) * time.Second,
-				RenewDeadline:   time.Duration(sm.config.RenewDeadline) * time.Second,
-				RetryPeriod:     time.Duration(sm.config.RetryPeriod) * time.Second,
-				Callbacks: leaderelection.LeaderCallbacks{
-					OnStartedLeading: func(ctx context.Context) {
-						err = sm.svcProcessor.ServicesWatcher(ctx, sm.svcProcessor.SyncServices)
-						if err != nil {
-							log.Error(err.Error())
-							panic("")
-						}
-					},
-					OnStoppedLeading: func() {
-						// we can do cleanup here
-						sm.mutex.Lock()
-						defer sm.mutex.Unlock()
-						log.Info("leader lost", "id", id)
-						sm.svcProcessor.Stop()
-
-						log.Error("lost leadership, restarting kube-vip")
-						panic("")
-					},
-					OnNewLeader: func(identity string) {
-						// we're notified when new leader elected
-						if identity == id {
-							// I just got the lock
-							return
-						}
-						log.Info("new leader elected", "id", identity)
-					},
-				},
-			})
-		} else {
-			log.Info("beginning watching services without leader election")
-			err = sm.svcProcessor.ServicesWatcher(ctx, sm.svcProcessor.SyncServices)
-			if err != nil {
-				log.Error("Cannot watch services", "err", err)
-			}
-		}
-	}
-
-	return nil
-}
-
-func (sm *Manager) cleanRoutes() error {
-	sm.mutex.Lock()
-	defer sm.mutex.Unlock()
-	routes, err := vip.ListRoutes(sm.config.RoutingTableID, sm.config.RoutingProtocol)
-	if err != nil {
-		return fmt.Errorf("error getting routes: %w", err)
-	}
-
-	for i := range routes {
-		found := false
-		if sm.config.EnableControlPlane {
-			found = (routes[i].Dst.IP.String() == sm.config.Address)
-		} else {
-			found = endpoints.CountRouteReferences(&routes[i], &sm.svcProcessor.ServiceInstances) > 0
-		}
-
-		if !found {
-			err = netlink.RouteDel(&(routes[i]))
-			if err != nil {
-				log.Error("[route] deletion", "route", routes[i], "err", err)
-			}
-			log.Debug("[route] deletion", "route", routes[i])
-		}
-
-	}
-	return nil
-}

pkg/manager/manager_test.go

@@ -0,0 +1,38 @@
+package manager
+
+import (
+	"testing"
+
+	"github.com/stretchr/testify/assert"
+)
+
+func TestNormalizeNodeName(t *testing.T) {
+	tests := []struct {
+		name     string
+		hostname string
+		expected string
+	}{
+		{
+			name:     "All lowercase hostname remains unchanged",
+			hostname: "worker-node-1",
+			expected: "worker-node-1",
+		},
+		{
+			name:     "Mixed case hostname is lowercased",
+			hostname: "Worker-Node-1",
+			expected: "worker-node-1",
+		},
+		{
+			name:     "All uppercase hostname is lowercased",
+			hostname: "MASTER-NODE",
+			expected: "master-node",
+		},
+	}
+
+	for _, tt := range tests {
+		t.Run(tt.name, func(t *testing.T) {
+			result := normalizeNodeName(tt.hostname)
+			assert.Equal(t, tt.expected, result, "The normalized node name did not match the expected RFC1123 compliant name")
+		})
+	}
+}

pkg/manager/manager_wireguard.go

@@ -1,122 +0,0 @@
-package manager
-
-import (
-	"context"
-	"time"
-
-	log "log/slog"
-
-	"github.com/kube-vip/kube-vip/pkg/wireguard"
-	metav1 "k8s.io/apimachinery/pkg/apis/meta/v1"
-	"k8s.io/client-go/tools/leaderelection"
-	"k8s.io/client-go/tools/leaderelection/resourcelock"
-)
-
-// Start will begin the Manager, which will start services and watch the configmap
-func (sm *Manager) startWireguard(id string) error {
-	var ns string
-	var err error
-
-	// use a Go context so we can tell the leaderelection code when we
-	// want to step down
-	ctx, cancel := context.WithCancel(context.Background())
-	defer cancel()
-	log.Info("reading wireguard peer configuration from Kubernetes secret")
-	s, err := sm.clientSet.CoreV1().Secrets(sm.config.Namespace).Get(ctx, "wireguard", metav1.GetOptions{})
-	if err != nil {
-		return err
-	}
-	// parse all the details needed for Wireguard
-	peerPublicKey := s.Data["peerPublicKey"]
-	peerEndpoint := s.Data["peerEndpoint"]
-	privateKey := s.Data["privateKey"]
-
-	// Configure the interface to join the Wireguard VPN
-	err = wireguard.ConfigureInterface(string(privateKey), string(peerPublicKey), string(peerEndpoint))
-	if err != nil {
-		return err
-	}
-
-	// Shutdown function that will wait on this signal, unless we call it ourselves
-	go func() {
-		<-sm.signalChan
-		log.Info("Received termination, signaling shutdown")
-
-		// Cancel the context, which will in turn cancel the leadership
-		cancel()
-	}()
-
-	ns, err = returnNameSpace()
-	if err != nil {
-		log.Warn("unable to auto-detect namespace", "dropping to", sm.config.Namespace)
-		ns = sm.config.Namespace
-	}
-
-	// Start a services watcher (all kube-vip pods will watch services), upon a new service
-	// a lock based upon that service is created that they will all leaderElection on
-	if sm.config.EnableServicesElection {
-		log.Info("beginning watching services, leaderelection will happen for every service")
-		err = sm.svcProcessor.StartServicesWatchForLeaderElection(ctx)
-		if err != nil {
-			return err
-		}
-	} else {
-
-		log.Info("beginning services leadership", "namespace", ns, "lock name", plunderLock, "id", id)
-		// we use the Lease lock type since edits to Leases are less common
-		// and fewer objects in the cluster watch "all Leases".
-		lock := &resourcelock.LeaseLock{
-			LeaseMeta: metav1.ObjectMeta{
-				Name:      plunderLock,
-				Namespace: ns,
-			},
-			Client: sm.clientSet.CoordinationV1(),
-			LockConfig: resourcelock.ResourceLockConfig{
-				Identity: id,
-			},
-		}
-
-		// start the leader election code loop
-		leaderelection.RunOrDie(ctx, leaderelection.LeaderElectionConfig{
-			Lock: lock,
-			// IMPORTANT: you MUST ensure that any code you have that
-			// is protected by the lease must terminate **before**
-			// you call cancel. Otherwise, you could have a background
-			// loop still running and another process could
-			// get elected before your background loop finished, violating
-			// the stated goal of the lease.
-			ReleaseOnCancel: true,
-			LeaseDuration:   time.Duration(sm.config.LeaseDuration) * time.Second,
-			RenewDeadline:   time.Duration(sm.config.RenewDeadline) * time.Second,
-			RetryPeriod:     time.Duration(sm.config.RetryPeriod) * time.Second,
-			Callbacks: leaderelection.LeaderCallbacks{
-				OnStartedLeading: func(ctx context.Context) {
-					err = sm.svcProcessor.ServicesWatcher(ctx, sm.svcProcessor.SyncServices)
-					if err != nil {
-						log.Error(err.Error())
-						panic("")
-					}
-				},
-				OnStoppedLeading: func() {
-					// we can do cleanup here
-					sm.mutex.Lock()
-					defer sm.mutex.Unlock()
-					log.Info("leader lost", "id", id)
-					sm.svcProcessor.Stop()
-
-					log.Error("lost leadership, restarting kube-vip")
-					panic("")
-				},
-				OnNewLeader: func(identity string) {
-					// we're notified when new leader elected
-					if identity == id {
-						// I just got the lock
-						return
-					}
-					log.Info("new leader elected", "id", identity)
-				},
-			},
-		})
-	}
-	return nil
-}

pkg/manager/node_labeling.go

@@ -1,73 +0,0 @@
-package manager
-
-import (
-	"context"
-	"encoding/json"
-	"fmt"
-
-	log "log/slog"
-
-	metav1 "k8s.io/apimachinery/pkg/apis/meta/v1"
-	"k8s.io/apimachinery/pkg/types"
-	"k8s.io/client-go/kubernetes"
-)
-
-const (
-	nodeLabelIndex    = "kube-vip.io/has-ip"
-	nodeLabelJSONPath = `kube-vip.io~1has-ip`
-)
-
-type patchStringLabel struct {
-	Op    string `json:"op"`
-	Path  string `json:"path"`
-	Value string `json:"value"`
-}
-
-// applyNodeLabel add/remove node label `kube-vip.io/has-ip=<VIP-Address>` to/from
-// the node where the virtual IP was added to/removed from.
-func applyNodeLabel(clientSet *kubernetes.Clientset, address, id, identity string) {
-	ctx := context.Background()
-	node, err := clientSet.CoreV1().Nodes().Get(ctx, id, metav1.GetOptions{})
-	if err != nil {
-		log.Error("can't query node labels", "node", id, "err", err)
-		return
-	}
-
-	log.Debug(fmt.Sprintf("node %s labels: %+v", id, node.Labels))
-
-	value, ok := node.Labels[nodeLabelIndex]
-	path := fmt.Sprintf("/metadata/labels/%s", nodeLabelJSONPath)
-	log.Debug(fmt.Sprintf("Received identity: %s - id: %s", identity, id))
-	if ok && value == address {
-		log.Debug(fmt.Sprintf("removing node label `has-ip=%s` on %s", address, id))
-		// Remove label
-		applyPatchLabels(ctx, clientSet, id, "remove", path, address)
-	} else {
-		log.Debug(fmt.Sprintf("setting node label `has-ip=%s` on %s", address, id))
-		// Append label
-		applyPatchLabels(ctx, clientSet, id, "add", path, address)
-	}
-}
-
-// applyPatchLabels add/remove node labels
-func applyPatchLabels(ctx context.Context, clientSet *kubernetes.Clientset,
-	name, operation, path, value string) {
-	patchLabels := []patchStringLabel{{
-		Op:    operation,
-		Path:  path,
-		Value: value,
-	}}
-	patchData, err := json.Marshal(patchLabels)
-	if err != nil {
-		log.Error("node patch marshaling failed", "err", err)
-		return
-	}
-	// patch node
-	node, err := clientSet.CoreV1().Nodes().Patch(ctx,
-		name, types.JSONPatchType, patchData, metav1.PatchOptions{})
-	if err != nil {
-		log.Error("node patching failed", "err", err)
-		return
-	}
-	log.Debug("updated", "node", name, "labels", node.Labels)
-}

pkg/manager/prom.go

@@ -1,15 +0,0 @@
-package manager
-
-import "github.com/prometheus/client_golang/prometheus"
-
-// PrometheusCollector defines a service watch event counter.
-func (sm *Manager) PrometheusCollector() []prometheus.Collector {
-	collectors := []prometheus.Collector{}
-	if sm.svcProcessor != nil {
-		collectors = append(collectors, sm.svcProcessor.CountServiceWatchEvent)
-	}
-	if sm.bgpServer != nil {
-		collectors = append(collectors, sm.bgpServer.BGPSessionInfoGauge)
-	}
-	return collectors
-}

pkg/manager/watch_annotations.go

@@ -10,13 +10,13 @@ import (
 
 	log "log/slog"
 
-	"github.com/kube-vip/kube-vip/pkg/bgp"
+	"github.com/kube-vip/kube-vip/pkg/kubevip"
+	"github.com/kube-vip/kube-vip/pkg/utils"
 
-	"github.com/davecgh/go-spew/spew"
 	v1 "k8s.io/api/core/v1"
-	apierrors "k8s.io/apimachinery/pkg/api/errors"
 	metav1 "k8s.io/apimachinery/pkg/apis/meta/v1"
 	"k8s.io/apimachinery/pkg/labels"
+	"k8s.io/client-go/kubernetes"
 	watchtools "k8s.io/client-go/tools/watch"
 
 	"k8s.io/apimachinery/pkg/watch"
@@ -25,18 +25,19 @@ import (
 
 // This file handles the watching of node annotations for configuration, it will exit once the annotations are
 // present
-func (sm *Manager) annotationsWatcher() error {
+func annotationsWatcher(ctx context.Context, clientSet,
+	rwClientSet kubernetes.Interface, config *kubevip.Config) error {
 	// Use a restartable watcher, as this should help in the event of etcd or timeout issues
-	log.Info("Kube-Vip is waiting for annotation prefix to be present on this node", "prefix", sm.config.Annotations)
+	log.Info("Kube-Vip is waiting for annotation prefix to be present on this node", "prefix", config.Annotations)
 
-	labelSelector := metav1.LabelSelector{MatchLabels: map[string]string{"kubernetes.io/hostname": sm.config.NodeName}}
+	labelSelector := metav1.LabelSelector{MatchLabels: map[string]string{"kubernetes.io/hostname": config.NodeName}}
 	listOptions := metav1.ListOptions{
 		LabelSelector: labels.Set(labelSelector.MatchLabels).String(),
 	}
 
 	// First we'll check the annotations for the node and if
 	// they aren't what are expected, we'll drop into the watch until they are
-	nodeList, err := sm.clientSet.CoreV1().Nodes().List(context.Background(), listOptions)
+	nodeList, err := clientSet.CoreV1().Nodes().List(ctx, listOptions)
 	if err != nil {
 		return err
 	}
@@ -45,11 +46,11 @@ func (sm *Manager) annotationsWatcher() error {
 	// there's probably bigger problems
 	node := nodeList.Items[0]
 
-	bgpConfig, bgpPeer, err := parseBgpAnnotations(sm.config.BGPConfig, &node, sm.config.Annotations)
+	bgpConfig, bgpPeer, err := parseBgpAnnotations(config.BGPConfig, &node, config.Annotations)
 	if err == nil {
 		// No error, the annotations already exist
-		sm.config.BGPConfig = bgpConfig
-		sm.config.BGPPeerConfig = bgpPeer
+		config.BGPConfig = bgpConfig
+		config.BGPPeerConfig = bgpPeer
 		return nil
 	}
 
@@ -57,30 +58,19 @@ func (sm *Manager) annotationsWatcher() error {
 	// they're as needed
 	log.Warn(err.Error())
 
-	// TODO, will need refactoring as part of rikatz work
-	rw, err := watchtools.NewRetryWatcherWithContext(context.TODO(), node.ResourceVersion, &cache.ListWatch{
+	rw, err := watchtools.NewRetryWatcherWithContext(ctx, node.ResourceVersion, &cache.ListWatch{
 		WatchFunc: func(_ metav1.ListOptions) (watch.Interface, error) {
-			return sm.rwClientSet.CoreV1().Nodes().Watch(context.Background(), listOptions)
+			return rwClientSet.CoreV1().Nodes().Watch(ctx, listOptions)
 		},
 	})
+
 	if err != nil {
 		return fmt.Errorf("error creating annotations watcher: %s", err.Error())
 	}
 
-	exitFunction := make(chan struct{})
-	go func() {
-		select {
-		case <-sm.shutdownChan:
-			log.Debug("[annotations] shutdown called")
-			// Stop the retry watcher
-			rw.Stop()
-			return
-		case <-exitFunction:
-			log.Debug("[annotations] function ending")
-			// Stop the retry watcher
-			rw.Stop()
-			return
-		}
+	defer func() {
+		rw.Stop()
+		log.Debug("[annotations] watcher stopped")
 	}()
 
 	ch := rw.ResultChan()
@@ -94,16 +84,17 @@ func (sm *Manager) annotationsWatcher() error {
 				return fmt.Errorf("unable to parse Kubernetes Node from Annotation watcher")
 			}
 
-			bgpConfig, bgpPeer, err := parseBgpAnnotations(sm.config.BGPConfig, node, sm.config.Annotations)
+			bgpConfig, bgpPeer, err := parseBgpAnnotations(config.BGPConfig, node, config.Annotations)
 			if err != nil {
 				log.Error(err.Error())
 				continue
 			}
 
-			sm.config.BGPConfig = bgpConfig
-			sm.config.BGPPeerConfig = bgpPeer
+			config.BGPConfig = bgpConfig
+			config.BGPPeerConfig = bgpPeer
 
-			rw.Stop()
+			log.Info("[annotations] exiting Annotations watcher - annotations found")
+			return nil
 		case watch.Deleted:
 			node, ok := event.Object.(*v1.Node)
 			if !ok {
@@ -116,24 +107,15 @@ func (sm *Manager) annotationsWatcher() error {
 			// Un-used
 		case watch.Error:
 			log.Error("Error attempting to watch Kubernetes Nodes")
-
-			// This round trip allows us to handle unstructured status
-			errObject := apierrors.FromObject(event.Object)
-			statusErr, ok := errObject.(*apierrors.StatusError)
-			if !ok {
-				log.Error(spew.Sprintf("Received an error which is not *metav1.Status but %#+v", event.Object))
-
-			}
-
-			status := statusErr.ErrStatus
-			log.Error(status.String())
+			log.Error("annotations watcher failed", "err", utils.WatchError(event.Object))
 		default:
 		}
 	}
-	close(exitFunction)
-	log.Info("Exiting Annotations watcher")
-	return nil
-
+	log.Info("[annotations] exiting annotations watcher")
+	if ctx.Err() != nil {
+		return nil
+	}
+	return utils.NewPanicError("annotations watcher channel closed unexpectedly")
 }
 
 // parseNodeAnnotations parses the annotations on the node and updates the configuration
@@ -151,12 +133,14 @@ func (sm *Manager) annotationsWatcher() error {
 // * `<info>` is the relevant information, such as `node-asn` or `peer-ip`
 // * `{{n}}` is the number of the peer, always starting with `0`
 // * kube-vip is only designed to manage one peer, just look for {{n}} == 0
-func parseBgpAnnotations(bgpConfig bgp.Config, node *v1.Node, prefix string) (bgp.Config, bgp.Peer, error) {
-	bgpPeer := bgp.Peer{}
+func parseBgpAnnotations(bgpConfig kubevip.BGPConfig, node *v1.Node, prefix string) (kubevip.BGPConfig, kubevip.BGPPeer, error) {
+	bgpPeer := kubevip.BGPPeer{}
+
+	prefix = regexp.QuoteMeta(prefix)
 
 	nodeASN := ""
+	regex := regexp.MustCompile(fmt.Sprintf("^%s/(bgp-peers-0-)?node-asn$", prefix))
 	for k, v := range node.Annotations {
-		regex := regexp.MustCompile(fmt.Sprintf("^%s/(bgp-peers-0-)?node-asn", prefix))
 		if regex.Match([]byte(k)) {
 			nodeASN = v
 		}
@@ -173,8 +157,8 @@ func parseBgpAnnotations(bgpConfig bgp.Config, node *v1.Node, prefix string) (bg
 	bgpConfig.AS = uint32(u64)
 
 	srcIP := ""
+	regex = regexp.MustCompile(fmt.Sprintf("^%s/(bgp-peers-0-)?src-ip$", prefix))
 	for k, v := range node.Annotations {
-		regex := regexp.MustCompile(fmt.Sprintf("^%s/(bgp-peers-0-)?src-ip", prefix))
 		if regex.Match([]byte(k)) {
 			srcIP = v
 		}
@@ -188,8 +172,8 @@ func parseBgpAnnotations(bgpConfig bgp.Config, node *v1.Node, prefix string) (bg
 	bgpConfig.RouterID, bgpConfig.SourceIP = srcIP, srcIP
 
 	peerASN := ""
+	regex = regexp.MustCompile(fmt.Sprintf("^%s/(bgp-peers-0-)?peer-asn$", prefix))
 	for k, v := range node.Annotations {
-		regex := regexp.MustCompile(fmt.Sprintf("^%s/(bgp-peers-0-)?peer-asn", prefix))
 		if regex.Match([]byte(k)) {
 			peerASN = v
 		}
@@ -206,8 +190,8 @@ func parseBgpAnnotations(bgpConfig bgp.Config, node *v1.Node, prefix string) (bg
 	bgpPeer.AS = uint32(u64)
 
 	peerIPString := ""
+	regex = regexp.MustCompile(fmt.Sprintf("^%s/(bgp-peers-[0-9]+-)?peer-ip$", prefix))
 	for k, v := range node.Annotations {
-		regex := regexp.MustCompile(fmt.Sprintf("^%s/(bgp-peers-[0-9]+-)?peer-ip", prefix))
 		if regex.Match([]byte(k)) {
 			peerIPString += v + ","
 		}
@@ -215,8 +199,13 @@ func parseBgpAnnotations(bgpConfig bgp.Config, node *v1.Node, prefix string) (bg
 	peerIPString = strings.TrimRight(peerIPString, ",")
 
 	peerIPs := strings.Split(peerIPString, ",")
+	if len(peerIPs) >= 1 && peerIPs[0] == "" || len(peerIPs) == 0 {
+		return bgpConfig, bgpPeer, fmt.Errorf("peer-ip value missing or empty")
+	}
 
-	bgpConfig.Peers = make([]bgp.Peer, 0, len(peerIPs))
+	bgpConfig.Peers = make([]kubevip.BGPPeer, 0, len(peerIPs))
+	regexPass := regexp.MustCompile(fmt.Sprintf("^%s/(bgp-peers-0-)?bgp-pass$", prefix))
+	regexMultiHop := regexp.MustCompile(fmt.Sprintf("^%s/(bgp-peers-0-)?peer-multi-hop$", prefix))
 	for _, peerIP := range peerIPs {
 		ipAddr := strings.TrimSpace(peerIP)
 
@@ -225,8 +214,7 @@ func parseBgpAnnotations(bgpConfig bgp.Config, node *v1.Node, prefix string) (bg
 			// Check if we're also expecting a password for this peer
 			base64BGPPassword := ""
 			for k, v := range node.Annotations {
-				regex := regexp.MustCompile(fmt.Sprintf("^%s/(bgp-peers-0-)?bgp-pass", prefix))
-				if regex.Match([]byte(k)) {
+				if regexPass.Match([]byte(k)) {
 					base64BGPPassword = v
 				}
 			}
@@ -239,6 +227,21 @@ func parseBgpAnnotations(bgpConfig bgp.Config, node *v1.Node, prefix string) (bg
 				// Set the password for each peer
 				bgpPeer.Password = string(decodedPassword)
 			}
+
+			// Check if multi-hop is enabled.
+			for k, v := range node.Annotations {
+				if regexMultiHop.MatchString(k) {
+					switch v {
+					case "true":
+						bgpPeer.MultiHop = true
+					case "false":
+						bgpPeer.MultiHop = false
+					default:
+						return bgpConfig, bgpPeer,
+							fmt.Errorf("invalid %q annotation value: %q, must be \"true\" or \"false\"", k, v)
+					}
+				}
+			}
 			bgpConfig.Peers = append(bgpConfig.Peers, bgpPeer)
 		}
 	}

pkg/manager/watcher_test.go

@@ -4,7 +4,7 @@ import (
 	"reflect"
 	"testing"
 
-	"github.com/kube-vip/kube-vip/pkg/bgp"
+	"github.com/kube-vip/kube-vip/pkg/kubevip"
 	"github.com/stretchr/testify/assert"
 	corev1 "k8s.io/api/core/v1"
 	metav1 "k8s.io/apimachinery/pkg/apis/meta/v1"
@@ -15,7 +15,7 @@ func TestParseBgpAnnotations(t *testing.T) {
 		ObjectMeta: metav1.ObjectMeta{Name: "test", Annotations: map[string]string{}},
 	}
 
-	bgpConfigBase := bgp.Config{
+	bgpConfigBase := kubevip.BGPConfig{
 		HoldTime:          15,
 		KeepaliveInterval: 5,
 	}
@@ -25,9 +25,11 @@ func TestParseBgpAnnotations(t *testing.T) {
 	}
 
 	node.Annotations = map[string]string{
-		"bgp/node-asn": "65000",
-		"bgp/peer-asn": "64000",
-		"bgp/src-ip":   "10.0.0.254",
+		"bgp/node-asn":       "65000",
+		"bgp/peer-asn":       "64000",
+		"bgp/src-ip":         "10.0.0.254",
+		"bgp/peer-ip":        "10.0.0.1",
+		"bgp/peer-multi-hop": "true",
 	}
 
 	bgpConfig, bgpPeer, err := parseBgpAnnotations(bgpConfigBase, node, "bgp")
@@ -38,15 +40,17 @@ func TestParseBgpAnnotations(t *testing.T) {
 	assert.Equal(t, uint32(65000), bgpConfig.AS, "bgpConfig.AS parsed incorrectly")
 	assert.Equal(t, uint32(64000), bgpPeer.AS, "bgpPeer.AS parsed incorrectly")
 	assert.Equal(t, "10.0.0.254", bgpConfig.RouterID, "bgpConfig.RouterID parsed incorrectly")
+	assert.Equal(t, true, bgpPeer.MultiHop, "bgpPeer.MultiHop parsed incorrectly")
 	assert.EqualValues(t, 15, bgpConfig.HoldTime, "base bgpConfig.HoldTime should not be overwritten")
 	assert.EqualValues(t, 5, bgpConfig.KeepaliveInterval, "base bgpConfig.KeepaliveInterval should not be overwritten")
 
-	node.Annotations = map[string]string{
-		"bgp/node-asn": "65000",
-		"bgp/peer-asn": "64000",
-		"bgp/src-ip":   "10.0.0.254",
-		"bgp/peer-ip":  "10.0.0.1,10.0.0.2,10.0.0.3",
-		"bgp/bgp-pass": "cGFzc3dvcmQ=", // password
+	node.Annotations = map[string]string{ //nolint:gosec
+		"bgp/node-asn":       "65000",
+		"bgp/peer-asn":       "64000",
+		"bgp/src-ip":         "10.0.0.254",
+		"bgp/peer-ip":        "10.0.0.1,10.0.0.2,10.0.0.3",
+		"bgp/bgp-pass":       "cGFzc3dvcmQ=", // dummy password for the test, gosec linter disabled
+		"bgp/peer-multi-hop": "true",
 	}
 
 	bgpConfig, bgpPeer, err = parseBgpAnnotations(bgpConfigBase, node, "bgp")
@@ -54,10 +58,10 @@ func TestParseBgpAnnotations(t *testing.T) {
 		t.Fatal("Parsing BGP annotations should return nil when minimum config is met")
 	}
 
-	bgpPeers := []bgp.Peer{
-		{Address: "10.0.0.1", AS: uint32(64000), Password: "password"},
-		{Address: "10.0.0.2", AS: uint32(64000), Password: "password"},
-		{Address: "10.0.0.3", AS: uint32(64000), Password: "password"},
+	bgpPeers := []kubevip.BGPPeer{
+		{Address: "10.0.0.1", AS: uint32(64000), Password: "password", MultiHop: true},
+		{Address: "10.0.0.2", AS: uint32(64000), Password: "password", MultiHop: true},
+		{Address: "10.0.0.3", AS: uint32(64000), Password: "password", MultiHop: true},
 	}
 	assert.Equal(t, bgpPeers, bgpConfig.Peers, "bgpConfig.Peers parsed incorrectly")
 	assert.Equal(t, "10.0.0.3", bgpPeer.Address, "bgpPeer.Address parsed incorrectly")
@@ -77,7 +81,7 @@ func TestParseNewBgpAnnotations(t *testing.T) {
 		ObjectMeta: metav1.ObjectMeta{Name: "test", Annotations: map[string]string{}},
 	}
 
-	bgpConfigBase := bgp.Config{
+	bgpConfigBase := kubevip.BGPConfig{
 		HoldTime:          15,
 		KeepaliveInterval: 5,
 	}
@@ -86,12 +90,12 @@ func TestParseNewBgpAnnotations(t *testing.T) {
 		t.Fatal("Parsing BGP annotations should return an error when no annotations exist")
 	}
 
-	node.Annotations = map[string]string{
+	node.Annotations = map[string]string{ //nolint:gosec
 		"bgp/bgp-peers-0-node-asn": "65000",
 		"bgp/bgp-peers-0-peer-asn": "64000",
 		"bgp/bgp-peers-0-peer-ip":  "10.0.0.1,10.0.0.2,10.0.0.3",
 		"bgp/bgp-peers-0-src-ip":   "10.0.0.254",
-		"bgp/bgp-peers-0-bgp-pass": "cGFzc3dvcmQ=", // password
+		"bgp/bgp-peers-0-bgp-pass": "cGFzc3dvcmQ=", // dummy password for the test, gosec linter disabled
 	}
 
 	bgpConfig, bgpPeer, err := parseBgpAnnotations(bgpConfigBase, node, "bgp")
@@ -99,7 +103,7 @@ func TestParseNewBgpAnnotations(t *testing.T) {
 		t.Fatalf("Parsing BGP annotations should return nil when minimum config is met [%v]", err)
 	}
 
-	bgpPeers := []bgp.Peer{
+	bgpPeers := []kubevip.BGPPeer{
 		{Address: "10.0.0.1", AS: uint32(64000), Password: "password"},
 		{Address: "10.0.0.2", AS: uint32(64000), Password: "password"},
 		{Address: "10.0.0.3", AS: uint32(64000), Password: "password"},
@@ -121,15 +125,15 @@ func Test_parseBgpAnnotations(t *testing.T) {
 	tests := []struct {
 		name    string
 		args    args
-		want    bgp.Config
-		want1   bgp.Peer
+		want    kubevip.BGPConfig
+		want1   kubevip.BGPPeer
 		wantErr bool
 	}{
 		// TODO: Add test cases.
 	}
 	for _, tt := range tests {
 		t.Run(tt.name, func(t *testing.T) {
-			got, got1, err := parseBgpAnnotations(bgp.Config{}, tt.args.node, tt.args.prefix)
+			got, got1, err := parseBgpAnnotations(kubevip.BGPConfig{}, tt.args.node, tt.args.prefix)
 			if (err != nil) != tt.wantErr {
 				t.Errorf("parseBgpAnnotations() error = %v, wantErr %v", err, tt.wantErr)
 				return

pkg/manager/worker/arp.go

@@ -0,0 +1,73 @@
+package worker
+
+import (
+	"context"
+	log "log/slog"
+	"sync"
+	"sync/atomic"
+
+	"github.com/kube-vip/kube-vip/pkg/arp"
+	"github.com/kube-vip/kube-vip/pkg/election"
+	"github.com/kube-vip/kube-vip/pkg/kubevip"
+	"github.com/kube-vip/kube-vip/pkg/lease"
+	"github.com/kube-vip/kube-vip/pkg/networkinterface"
+	"github.com/kube-vip/kube-vip/pkg/node"
+	"github.com/kube-vip/kube-vip/pkg/route"
+	"github.com/kube-vip/kube-vip/pkg/services"
+	"k8s.io/client-go/kubernetes"
+)
+
+type ARP struct {
+	Common
+}
+
+func NewARP(arpMgr *arp.Manager, intfMgr *networkinterface.Manager,
+	config *kubevip.Config, closing *atomic.Bool, killFunc func(),
+	svcProcessor *services.Processor, mutex *sync.Mutex, clientSet *kubernetes.Clientset,
+	electionMgr *election.Manager, leaseMgr *lease.Manager, routeMgr *route.Manager,
+	nodeLabelMgr node.Labeler,
+) *ARP {
+	return &ARP{
+		Common: *newCommon(arpMgr, intfMgr, config, closing, killFunc,
+			svcProcessor, mutex, clientSet, electionMgr, leaseMgr, routeMgr,
+			nodeLabelMgr),
+	}
+}
+
+func (a *ARP) Configure(ctx context.Context, wg *sync.WaitGroup) error {
+	log.Info("Start ARP/NDP advertisement Global")
+	wg.Go(func() {
+		a.arpMgr.StartAdvertisement(ctx, a.killFunc)
+	})
+	return nil
+}
+
+func (a *ARP) StartControlPlane(ctx context.Context, electionManager *election.Manager) {
+	err := a.cpCluster.StartCluster(ctx, a.config, electionManager, nil, a.leaseMgr, a.killFunc)
+	if err != nil {
+		log.Error("starting control plane", "err", err)
+	}
+
+	// Trigger the shutdown of this manager instance
+	a.killFunc()
+}
+
+func (a *ARP) ConfigureServices() {
+}
+
+func (a *ARP) StartServices(ctx context.Context) error {
+	// Start a services watcher (all kube-vip pods will watch services), upon a new service
+	// a lock based upon that service is created that they will all leaderElection on
+	if a.config.EnableServicesElection {
+		if err := a.PerServiceLeader(ctx, false); err != nil {
+			return err
+		}
+	} else {
+		a.GlobalLeader(ctx, a.config.ServicesLeaseName)
+	}
+	return nil
+}
+
+func (a *ARP) Name() string {
+	return "ARP"
+}

pkg/manager/worker/bgp.go

@@ -0,0 +1,130 @@
+package worker
+
+import (
+	"context"
+	"fmt"
+	log "log/slog"
+	"net"
+	"strconv"
+	"sync"
+	"sync/atomic"
+
+	"github.com/kube-vip/kube-vip/pkg/arp"
+	"github.com/kube-vip/kube-vip/pkg/bgp"
+	"github.com/kube-vip/kube-vip/pkg/election"
+	"github.com/kube-vip/kube-vip/pkg/kubevip"
+	"github.com/kube-vip/kube-vip/pkg/lease"
+	"github.com/kube-vip/kube-vip/pkg/metrics"
+	"github.com/kube-vip/kube-vip/pkg/networkinterface"
+	"github.com/kube-vip/kube-vip/pkg/node"
+	"github.com/kube-vip/kube-vip/pkg/route"
+	"github.com/kube-vip/kube-vip/pkg/services"
+	api "github.com/osrg/gobgp/v4/api"
+	"github.com/osrg/gobgp/v4/pkg/apiutil"
+
+	"github.com/prometheus/client_golang/prometheus"
+	"k8s.io/client-go/kubernetes"
+)
+
+type BGP struct {
+	Common
+	bgpServer *bgp.Server
+}
+
+func NewBGP(arpMgr *arp.Manager, intfMgr *networkinterface.Manager,
+	config *kubevip.Config, closing *atomic.Bool, killFunc func(),
+	svcProcessor *services.Processor, mutex *sync.Mutex, clientSet *kubernetes.Clientset,
+	bgpServer *bgp.Server, electionMgr *election.Manager, leaseMgr *lease.Manager,
+	routeMgr *route.Manager, nodeLabelMgr node.Labeler) *BGP {
+	return &BGP{
+		Common: *newCommon(arpMgr, intfMgr, config, closing, killFunc,
+			svcProcessor, mutex, clientSet, electionMgr, leaseMgr, routeMgr,
+			nodeLabelMgr),
+		bgpServer: bgpServer,
+	}
+}
+
+func (b *BGP) Configure(ctx context.Context, _ *sync.WaitGroup) error {
+	var err error
+	if b.bgpServer == nil {
+		b.bgpServer, err = bgp.NewBGPServer(b.config.BGPConfig, log.Level(b.config.Logging))
+		if err != nil {
+			return fmt.Errorf("creating BGP server: %w", err)
+		}
+	}
+
+	log.Info("Starting the BGP server to advertise VIP routes to BGP peers")
+	if err := b.bgpServer.Start(ctx, func(p *apiutil.WatchEventMessage_PeerEvent) {
+		if p.Type != apiutil.PEER_EVENT_STATE {
+			return
+		}
+
+		ipaddr := p.Peer.State.NeighborAddress.String()
+
+		port := 179
+		peerDescription := net.JoinHostPort(ipaddr, strconv.Itoa(port))
+
+		for stateName, stateValue := range api.PeerState_SessionState_value {
+			metricValue := 0.0
+			if int(p.Peer.State.SessionState) == int(stateValue)-1 {
+
+				metricValue = 1
+			}
+
+			metrics.BGPSessionInfoGauge.With(prometheus.Labels{
+				"state": stateName,
+				"peer":  peerDescription,
+			}).Set(metricValue)
+		}
+	}); err != nil {
+		return fmt.Errorf("starting BGP server: %w", err)
+	}
+
+	return nil
+}
+
+func (b *BGP) Cleanup() {
+	// Defer a function to check if the bgpServer has been created and if so attempt to close it
+	if b.bgpServer != nil {
+		b.bgpServer.Close()
+	}
+}
+
+func (b *BGP) StartControlPlane(ctx context.Context, electionManager *election.Manager) {
+	var err error
+	if b.config.EnableLeaderElection {
+		err = b.cpCluster.StartCluster(ctx, b.config, electionManager, b.bgpServer, b.leaseMgr, b.killFunc)
+	} else {
+		err = b.cpCluster.StartVipService(ctx, b.config, electionManager, b.bgpServer, b.killFunc)
+	}
+	if err != nil {
+		log.Error("Control Plane", "err", err)
+		// Trigger the shutdown of this manager instance
+		b.killFunc()
+	}
+}
+
+func (b *BGP) ConfigureServices() {
+	// No configuration required
+}
+
+func (b *BGP) StartServices(ctx context.Context) error {
+	if b.config.EnableServicesElection {
+		if err := b.PerServiceLeader(ctx, false); err != nil {
+			return err
+		}
+	} else {
+		if err := b.ServicesNoLeader(ctx); err != nil {
+			return err
+		}
+	}
+	return nil
+}
+
+func (b *BGP) ServicesGlobalLeader(ctx context.Context, id string) {
+	// NOT IMPLEMENTED
+}
+
+func (b *BGP) Name() string {
+	return "BGP"
+}

pkg/manager/worker/common.go

@@ -0,0 +1,261 @@
+package worker
+
+import (
+	"context"
+	"fmt"
+	log "log/slog"
+	"sync"
+	"sync/atomic"
+
+	"github.com/kube-vip/kube-vip/pkg/arp"
+	"github.com/kube-vip/kube-vip/pkg/cluster"
+	"github.com/kube-vip/kube-vip/pkg/election"
+	"github.com/kube-vip/kube-vip/pkg/kubevip"
+	"github.com/kube-vip/kube-vip/pkg/lease"
+	"github.com/kube-vip/kube-vip/pkg/metrics"
+	"github.com/kube-vip/kube-vip/pkg/networkinterface"
+	"github.com/kube-vip/kube-vip/pkg/node"
+	"github.com/kube-vip/kube-vip/pkg/route"
+	"github.com/kube-vip/kube-vip/pkg/services"
+	"k8s.io/client-go/kubernetes"
+)
+
+type Common struct {
+	arpMgr       *arp.Manager
+	cpCluster    *cluster.Cluster
+	intfMgr      *networkinterface.Manager
+	config       *kubevip.Config
+	closing      *atomic.Bool
+	killFunc     func()
+	svcProcessor *services.Processor
+	mutex        *sync.Mutex
+	clientSet    *kubernetes.Clientset
+	electionMgr  *election.Manager
+	leaseMgr     *lease.Manager
+	routeMgr     *route.Manager
+	nodeLabelMgr node.Labeler
+}
+
+func newCommon(arpMgr *arp.Manager, intfMgr *networkinterface.Manager,
+	config *kubevip.Config, closing *atomic.Bool, killFunc func(),
+	svcProcessor *services.Processor, mutex *sync.Mutex, clientSet *kubernetes.Clientset,
+	electionMgr *election.Manager, leaseMgr *lease.Manager, routeMgr *route.Manager,
+	nodeLabelMgr node.Labeler) *Common {
+	return &Common{
+		arpMgr:       arpMgr,
+		intfMgr:      intfMgr,
+		config:       config,
+		closing:      closing,
+		killFunc:     killFunc,
+		svcProcessor: svcProcessor,
+		mutex:        mutex,
+		clientSet:    clientSet,
+		electionMgr:  electionMgr,
+		leaseMgr:     leaseMgr,
+		routeMgr:     routeMgr,
+		nodeLabelMgr: nodeLabelMgr,
+	}
+}
+
+func (c *Common) InitControlPlane() error {
+	var err error
+	c.cpCluster, err = cluster.InitCluster(c.config, false, c.intfMgr, c.arpMgr, c.routeMgr, c.nodeLabelMgr)
+	if err != nil {
+		return fmt.Errorf("cluster initialization error: %w", err)
+	}
+	return nil
+}
+
+func (c *Common) PerServiceLeader(ctx context.Context, forcedOnly bool) error {
+	if forcedOnly {
+		log.Info(fmt.Sprintf("beginning watching services, leaderelection will happen for services annotated with '%s = \"true\"'", kubevip.ForcePerServiceElection))
+	} else {
+		log.Info("beginning watching services, leaderelection will happen for every service")
+	}
+
+	err := c.svcProcessor.StartServicesWatchForLeaderElection(ctx, forcedOnly)
+	if err != nil {
+		return err
+	}
+	return nil
+}
+
+func (c *Common) GlobalLeader(ctx context.Context, leaseName string) {
+	wg := sync.WaitGroup{}
+	defer wg.Wait()
+
+	servicesCtx, servicesCtxCancel := context.WithCancel(ctx)
+	defer servicesCtxCancel()
+
+	if c.config.PerServiceElectionOnDemand {
+		wg.Go(func() {
+			if err := c.PerServiceLeader(servicesCtx, true); err != nil {
+				log.Error("per-service leader election failed with", "error", err)
+				servicesCtxCancel()
+			}
+		})
+	}
+
+	c.runGlobalElection(servicesCtx, c, leaseName, c.config, c.electionMgr)
+}
+
+func (c *Common) ServicesNoLeader(ctx context.Context) error {
+	wg := sync.WaitGroup{}
+	defer wg.Wait()
+
+	servicesCtx, servicesCtxCancel := context.WithCancel(ctx)
+	defer servicesCtxCancel()
+
+	if c.config.PerServiceElectionOnDemand {
+		wg.Go(func() {
+			if err := c.PerServiceLeader(servicesCtx, true); err != nil {
+				log.Error("per-service leader election failed with", "error", err)
+				servicesCtxCancel()
+			}
+		})
+	}
+
+	log.Info("beginning watching services without leader election")
+	err := c.svcProcessor.ServicesWatcher(servicesCtx, services.NewCallback(c.svcProcessor.SyncServices, false), false)
+	if err != nil {
+		return fmt.Errorf("error while watching services: %w", err)
+	}
+	return nil
+}
+
+func (c *Common) Cleanup() {
+	// NOT IMPLEMENTED
+}
+
+func (c *Common) OnStartedLeading(ctx context.Context) {
+	err := c.svcProcessor.ServicesWatcher(ctx, services.NewCallback(c.svcProcessor.SyncServices, false), false)
+	if err != nil {
+		log.Error("service watcher", "err", err)
+		c.killFunc()
+	}
+}
+
+func (c *Common) OnStoppedLeading() {
+	// we can do cleanup here
+	c.mutex.Lock()
+	defer c.mutex.Unlock()
+	log.Info("leader lost", "former leader", c.config.NodeName)
+	c.svcProcessor.Stop()
+
+	log.Error("lost services leadership, restarting kube-vip")
+	c.killFunc()
+}
+
+func (c *Common) OnNewLeader(identity string) {
+	if identity == c.config.NodeName {
+		// I just got the lock
+		return
+	}
+	log.Info("new leader elected", "new leader", identity)
+}
+
+func (c *Common) runGlobalElection(ctx context.Context, a election.Actions, leaseName string,
+	config *kubevip.Config, electionManager *election.Manager) {
+
+	log.Debug("starting global election")
+	ns, leaseName := lease.NamespaceName(leaseName, config)
+
+	leaseID := lease.NewID(config.LeaderElectionType, ns, leaseName)
+	objectName := lease.ObjectName(leaseID, "svcs0")
+
+	// objLease, isNew, isSharedLease := c.leaseMgr.Add(leaseID, objectName)
+
+	objLease := c.leaseMgr.Add(ctx, leaseID)
+	isNew := objLease.Add(objectName)
+
+	// this service was already processed so we do not need to do anything
+	if !isNew {
+		log.Debug("this election was already done, waiting for it to finish", "lease", c.config.ServicesLeaseName)
+		// Wait for either the service context or lease context to be done
+		select {
+		case <-ctx.Done():
+			// Service was deleted
+			c.leaseMgr.Delete(leaseID, objectName, objLease)
+		case <-objLease.Ctx.Done():
+			// Leader election ended (leadership lost or context cancelled)
+		}
+		return
+	}
+
+	objLease.Lock()
+
+	defer func() {
+		objLease.Unlock()
+	}()
+
+	if objLease.Elected.Load() {
+		objLease.Unlock()
+		log.Debug("this election was already done, shared lease", "lease", leaseID.Name())
+
+		// wait for leader election to start or context to be done
+		select {
+		case <-objLease.Started:
+		case <-objLease.Ctx.Done():
+			// Lease was cancelled (e.g., leader election ended), return immediately
+			// This allows the restart loop to create a fresh lease
+			log.Debug("lease context cancelled before leader election started", "lease", leaseID.Name())
+			return
+		}
+
+		a.OnStartedLeading(objLease.Ctx)
+
+		log.Debug("waiting for lease to finish", "lease", leaseID.Name())
+		// wait for leaderelection to be finished
+		<-objLease.Ctx.Done()
+
+		// we can do cleanup here
+		a.OnStoppedLeading()
+
+		log.Error("lost leadership, restarting kube-vip", "lease", leaseID.Name())
+		c.killFunc()
+
+		return
+	}
+
+	// For new leases (not shared), ensure cleanup when the leader election ends
+	// This is critical for the restartable service watcher to be able to restart
+	// the leader election after leadership loss
+	defer func() {
+		// Delete the lease from the manager so subsequent calls can create a fresh lease
+		// This handles the case where leader election ends due to:
+		// 1. Leadership loss (e.g., network timeout)
+		// 2. Context cancellation
+		// 3. Any other reason RunOrDie returns
+		c.leaseMgr.Delete(leaseID, objectName, objLease)
+	}()
+
+	wg := sync.WaitGroup{}
+	defer wg.Wait()
+
+	run := &election.RunConfig{
+		Config:           config,
+		LeaseID:          leaseID,
+		LeaseAnnotations: map[string]string{},
+		Mgr:              electionManager,
+		OnStartedLeading: func(ctx context.Context) {
+			wg.Go(func() {
+				objLease.Elected.Store(true)
+				objLease.Unlock()
+				close(objLease.Started)
+				a.OnStartedLeading(ctx)
+				metrics.LeaderTransitionsTotal.WithLabelValues(leaseID.Name()).Inc()
+				metrics.IsLeader.WithLabelValues(config.NodeName, leaseID.Name()).Set(1)
+			})
+		},
+		OnStoppedLeading: func() {
+			objLease.Elected.Store(false)
+			a.OnStoppedLeading()
+			metrics.IsLeader.WithLabelValues(config.NodeName, leaseID.Name()).Set(0)
+		},
+		OnNewLeader: a.OnNewLeader,
+	}
+
+	if err := election.RunOrDie(ctx, run, config); err != nil {
+		log.Error("leaderelection failed", "err", err, "id", config.NodeName, "name", leaseID.Name())
+	}
+}

pkg/manager/worker/table.go

@@ -0,0 +1,116 @@
+package worker
+
+import (
+	"context"
+	"fmt"
+	log "log/slog"
+	"sync"
+	"sync/atomic"
+	"time"
+
+	"github.com/kube-vip/kube-vip/pkg/arp"
+	"github.com/kube-vip/kube-vip/pkg/election"
+	"github.com/kube-vip/kube-vip/pkg/kubevip"
+	"github.com/kube-vip/kube-vip/pkg/lease"
+	"github.com/kube-vip/kube-vip/pkg/networkinterface"
+	"github.com/kube-vip/kube-vip/pkg/node"
+	"github.com/kube-vip/kube-vip/pkg/route"
+	"github.com/kube-vip/kube-vip/pkg/services"
+	"github.com/kube-vip/kube-vip/pkg/vip"
+	"github.com/vishvananda/netlink"
+	"k8s.io/client-go/kubernetes"
+)
+
+type Table struct {
+	Common
+}
+
+func NewTable(arpMgr *arp.Manager, intfMgr *networkinterface.Manager,
+	config *kubevip.Config, closing *atomic.Bool, killFUnc func(),
+	svcProcessor *services.Processor, mutex *sync.Mutex, clientSet *kubernetes.Clientset,
+	electionMgr *election.Manager, leaseMgr *lease.Manager, routeMgr *route.Manager,
+	nodeLabelMgr node.Labeler) *Table {
+	return &Table{
+		Common: *newCommon(arpMgr, intfMgr, config, closing, killFUnc,
+			svcProcessor, mutex, clientSet, electionMgr, leaseMgr, routeMgr,
+			nodeLabelMgr),
+	}
+}
+
+func (t *Table) Configure(ctx context.Context, wg *sync.WaitGroup) error {
+	log.Info("destination for routes", "table", t.config.RoutingTableID, "protocol", t.config.RoutingProtocol)
+
+	if t.config.CleanRoutingTable {
+		wg.Go(func() {
+			// we assume that after 10s all services should be configured so we can delete redundant routes
+			time.Sleep(time.Second * 10)
+			if err := t.cleanRoutes(); err != nil {
+				log.Error("error checking for old routes", "err", err)
+			}
+		})
+	}
+	return nil
+}
+
+func (t *Table) StartControlPlane(ctx context.Context, electionManager *election.Manager) {
+	if err := t.cpCluster.StartVipService(ctx, t.config, electionManager, nil, t.killFunc); err != nil {
+		log.Error("Control Plane", "err", err)
+		// Trigger the shutdown of this manager instance
+		t.killFunc()
+	} else {
+		log.Debug("start VipServer for cluster manager successful")
+	}
+}
+
+func (t *Table) ConfigureServices() {
+	// No configuration required
+}
+
+func (t *Table) StartServices(ctx context.Context) error {
+	log.Debug("starting Services")
+
+	if t.config.EnableServicesElection {
+		if err := t.PerServiceLeader(ctx, false); err != nil {
+			return err
+		}
+	} else if t.config.EnableLeaderElection {
+		t.GlobalLeader(ctx, t.config.ServicesLeaseName)
+	} else {
+		if err := t.ServicesNoLeader(ctx); err != nil {
+			return err
+		}
+	}
+	return nil
+}
+
+func (t *Table) Name() string {
+	return "Routing Table"
+}
+
+func (t *Table) cleanRoutes() error {
+	t.mutex.Lock()
+	defer t.mutex.Unlock()
+	routes, err := vip.ListRoutes(t.config.RoutingTableID, t.config.RoutingProtocol)
+	if err != nil {
+		return fmt.Errorf("error getting routes: %w", err)
+	}
+
+	for i := range routes {
+		found := false
+		if t.config.EnableControlPlane {
+			found = (routes[i].Dst.IP.String() == t.config.Address)
+		} else {
+			found = t.routeMgr.Check(vip.NetlinkHash(&(routes[i])))
+		}
+
+		if !found {
+			err = netlink.RouteDel(&(routes[i]))
+			if err != nil {
+				log.Error("[route] deletion", "route", routes[i], "err", err)
+			}
+			log.Debug("[route] deletion", "route", routes[i])
+		}
+
+	}
+	return nil
+}

pkg/manager/worker/wireguard.go

@@ -0,0 +1,278 @@
+package worker
+
+import (
+	"context"
+	"fmt"
+	log "log/slog"
+	"os"
+	"sync"
+	"sync/atomic"
+
+	"github.com/kube-vip/kube-vip/pkg/arp"
+	"github.com/kube-vip/kube-vip/pkg/election"
+	"github.com/kube-vip/kube-vip/pkg/endpoints/providers"
+	"github.com/kube-vip/kube-vip/pkg/kubevip"
+	"github.com/kube-vip/kube-vip/pkg/lease"
+	"github.com/kube-vip/kube-vip/pkg/networkinterface"
+	"github.com/kube-vip/kube-vip/pkg/nftables"
+	"github.com/kube-vip/kube-vip/pkg/node"
+	"github.com/kube-vip/kube-vip/pkg/route"
+	"github.com/kube-vip/kube-vip/pkg/services"
+	"github.com/kube-vip/kube-vip/pkg/sysctl"
+	"github.com/kube-vip/kube-vip/pkg/utils"
+	"github.com/kube-vip/kube-vip/pkg/wireguard"
+	v1 "k8s.io/api/core/v1"
+	metav1 "k8s.io/apimachinery/pkg/apis/meta/v1"
+	"k8s.io/apimachinery/pkg/watch"
+	"k8s.io/client-go/kubernetes"
+)
+
+type WireGuard struct {
+	Common
+	tunnelMgr           *wireguard.TunnelManager
+	kubeAPIHost         string
+	kubeAPIPort         string
+	endpointWatcherCtx  context.Context
+	endpointWatcherStop context.CancelFunc
+	endpointWatcherWg   sync.WaitGroup
+}
+
+func NewWireGuard(arpMgr *arp.Manager, intfMgr *networkinterface.Manager,
+	config *kubevip.Config, closing *atomic.Bool, killFUnc func(),
+	svcProcessor *services.Processor, mutex *sync.Mutex, clientSet *kubernetes.Clientset,
+	electionMgr *election.Manager, leaseMgr *lease.Manager, routeMgr *route.Manager,
+	nodeLabelMgr node.Labeler) *WireGuard {
+	return &WireGuard{
+		Common: *newCommon(arpMgr, intfMgr, config, closing, killFUnc,
+			svcProcessor, mutex, clientSet, electionMgr, leaseMgr, routeMgr,
+			nodeLabelMgr),
+	}
+}
+
+func (w *WireGuard) Configure(ctx context.Context, _ *sync.WaitGroup) error {
+	log.Info("reading wireguard tunnel configurations from Kubernetes secret")
+	tunnelMgr := wireguard.NewTunnelManager()
+
+	err := tunnelMgr.LoadConfigurationsFromSecret(ctx, w.clientSet, w.config.Namespace, "wireguard")
+	if err != nil {
+		return fmt.Errorf("failed to load WireGuard tunnel configurations: %w", err)
+	}
+
+	// Clean up any stale resources from previous runs (crash recovery for hostNetwork: true)
+	// Must be called AFTER loading configs so we know which interfaces/ports to clean
+	if err := tunnelMgr.CleanupStaleResources(); err != nil {
+		log.Warn("failed to cleanup stale resources", "err", err)
+		// Continue anyway - the cleanup is best-effort
+	}
+
+	if _, err := sysctl.EnableProcSys("/proc/sys/net/ipv4/conf/all/src_valid_mark"); err != nil {
+		return fmt.Errorf("net.ipv4.conf.all.src_valid_mark is disabled and could not be enabled %w", err)
+	}
+	if _, err := sysctl.EnableProcSys("/proc/sys/net/ipv4/conf/all/route_localnet"); err != nil {
+		return fmt.Errorf("net.ipv4.conf.all.route_localnet is disabled and could not be enabled %w", err)
+	}
+
+	w.tunnelMgr = tunnelMgr
+	configuredVIPs := tunnelMgr.ListConfiguredTunnels()
+	log.Info("loaded WireGuard tunnel configurations", "vips", configuredVIPs)
+
+	return nil
+}
+
+func (w *WireGuard) InitControlPlane() error {
+	// Get Kubernetes service IP and port from environment
+	w.kubeAPIHost = os.Getenv("KUBERNETES_SERVICE_HOST")
+	w.kubeAPIPort = os.Getenv("KUBERNETES_SERVICE_PORT_HTTPS")
+	if w.kubeAPIHost == "" || w.kubeAPIPort == "" {
+		return fmt.Errorf("KUBERNETES_SERVICE_HOST or KUBERNETES_SERVICE_PORT_HTTPS not set")
+	}
+	return nil
+}
+
+func (w *WireGuard) StartControlPlane(ctx context.Context, electionManager *election.Manager) {
+	if !w.tunnelMgr.HasConfigForVIP(w.config.VIP) {
+		log.Error("no WireGuard tunnel configuration found for control plane VIP", "vip", w.config.VIP)
+		return
+	}
+	w.runGlobalElection(ctx, w, w.config.LeaseName, w.config, electionManager)
+}
+
+func (w *WireGuard) ConfigureServices() {
+	w.svcProcessor.TunnelMgr = w.tunnelMgr
+}
+
+func (w *WireGuard) StartServices(ctx context.Context) error {
+	if w.config.EnableServicesElection {
+		log.Info("beginning watching services, leaderelection will happen for every service")
+		err := w.svcProcessor.StartServicesWatchForLeaderElection(ctx, false)
+		if err != nil {
+			return err
+		}
+	}
+	return nil
+}
+
+func (w *WireGuard) Name() string {
+	return "WireGuard"
+}
+
+func (w *WireGuard) OnStartedLeading(ctx context.Context) {
+	// Bring up the WireGuard tunnel for control plane VIP
+	err := w.tunnelMgr.BringUpTunnelForVIP(w.config.VIP)
+	if err != nil {
+		log.Error("could not start wireguard tunnel for control plane", "vip", w.config.VIP, "err", err)
+		_ = w.tunnelMgr.TearDownTunnelForVIP(w.config.VIP)
+		w.killFunc()
+		return
+	}
+
+	// Get the tunnel to access its configuration
+	wg := w.tunnelMgr.GetTunnelForVIP(w.config.VIP)
+	if wg == nil {
+		log.Error("failed to get wireguard tunnel after bringing up", "vip", w.config.VIP)
+		w.killFunc()
+		return
+	}
+
+	tunnelConfig := w.tunnelMgr.GetConfigForVIP(w.config.VIP)
+	if tunnelConfig == nil {
+		log.Error("failed to get tunnel configuration", "vip", w.config.VIP)
+		_ = w.tunnelMgr.TearDownTunnelForVIP(w.config.VIP)
+		w.killFunc()
+		return
+	}
+
+	// Start endpoint watcher - DNAT rules will be applied when endpoints arrive
+	w.endpointWatcherCtx, w.endpointWatcherStop = context.WithCancel(ctx)
+	w.endpointWatcherWg.Go(func() {
+		w.watchKubernetesEndpoints(w.endpointWatcherCtx, tunnelConfig)
+	})
+
+	if w.config.EnableServices && !w.config.EnableServicesElection {
+		if err := w.svcProcessor.ServicesWatcher(ctx, services.NewCallback(w.svcProcessor.SyncServices, false), false); err != nil {
+			log.Error("failed to start services watcher", "err", err)
+		}
+	}
+}
+
+// watchKubernetesEndpoints watches the kubernetes service EndpointSlices for changes
+// and updates the DNAT rules when API server endpoints change (e.g., when an API server goes down)
+func (w *WireGuard) watchKubernetesEndpoints(ctx context.Context, tunnelConfig *wireguard.TunnelConfig) {
+	log.Info("starting kubernetes endpoint watcher for control plane")
+
+	kubeSvc := &v1.Service{
+		ObjectMeta: metav1.ObjectMeta{
+			Name:      "kubernetes",
+			Namespace: "default",
+		},
+	}
+
+	provider := providers.NewEndpointslices()
+	rw, err := provider.CreateRetryWatcher(ctx, w.clientSet, kubeSvc)
+	if err != nil {
+		log.Error("failed to create kubernetes endpoint watcher", "err", err)
+		return
+	}
+	defer rw.Stop()
+
+	for event := range rw.ResultChan() {
+		select {
+		case <-ctx.Done():
+			log.Info("kubernetes endpoint watcher stopped")
+			return
+		default:
+		}
+
+		switch event.Type {
+		case watch.Added, watch.Modified, watch.Deleted:
+			if err := provider.LoadObject(event.Object, func() {}); err != nil {
+				log.Error("failed to load endpoint object", "err", err)
+				continue
+			}
+			endpoints, _ := provider.GetAllEndpoints()
+			log.Info("kubernetes endpoints changed, updating DNAT rules", "eventType", event.Type, "endpoints", endpoints)
+			if err := w.updateControlPlaneDNAT(tunnelConfig, endpoints); err != nil {
+				log.Error("failed to update control plane DNAT rules", "err", err)
+			}
+		case watch.Error:
+			log.Warn("kubernetes endpoint watch error", "event", event)
+		}
+	}
+}
+
+// updateControlPlaneDNAT updates the DNAT rules for the control plane with the given endpoints
+func (w *WireGuard) updateControlPlaneDNAT(tunnelConfig *wireguard.TunnelConfig, endpoints []string) error {
+	if len(endpoints) == 0 {
+		log.Warn("no kubernetes API server endpoints available")
+		// Don't delete rules - keep routing to last known endpoints
+		return nil
+	}
+
+	// Build targets with default port 6443
+	targets := make([]nftables.DNATTarget, len(endpoints))
+	for i, ep := range endpoints {
+		targets[i] = nftables.DNATTarget{IP: ep, Port: 6443}
+	}
+
+	vipIP := utils.StripCIDR(w.config.VIP)
+
+	err := nftables.ApplyDNAT(
+		tunnelConfig.InterfaceName,
+		vipIP,
+		6443,
+		targets,
+		"controlplane",
+		v1.ProtocolTCP,
+		false,
+		tunnelConfig.ListenPort,
+	)
+	if err != nil {
+		return fmt.Errorf("failed to apply updated DNAT rule: %w", err)
+	}
+
+	log.Info("control plane DNAT rules updated", "targetCount", len(targets))
+	return nil
+}
+
+func (w *WireGuard) OnStoppedLeading() {
+	// we can do cleanup here
+	w.mutex.Lock()
+	defer w.mutex.Unlock()
+	log.Info("leader lost", "id", w.config.NodeName)
+
+	// Stop the kubernetes endpoint watcher and wait for it to finish
+	if w.endpointWatcherStop != nil {
+		w.endpointWatcherStop()
+		w.endpointWatcherWg.Wait()
+	}
+
+	log.Info("deleting nftables DNAT chains")
+	err := nftables.DeleteIngressChains(false, "controlplane")
+	if err != nil {
+		log.Error("could not delete DNAT ingress chains", "err", err)
+	} else {
+		log.Info("nftables DNAT chains deleted successfully")
+	}
+
+	// Tear down all tunnels (control plane + services)
+	err = w.tunnelMgr.TearDownAllTunnels()
+	if err != nil {
+		log.Error("failed to tear down tunnels", "err", err)
+	}
+	if w.config.EnableServices && !w.config.EnableServicesElection {
+		w.svcProcessor.Stop()
+	}
+	log.Error("lost control plane leadership, restarting kube-vip")
+	w.killFunc()
+}
+
+func (w *WireGuard) OnNewLeader(identity string) {
+	// we're notified when new leader elected
+	if identity == w.config.NodeName {
+		// I just got the lock
+		return
+	}
+	// safety check - tear down tunnel if we're not the leader
+	_ = w.tunnelMgr.TearDownTunnelForVIP(w.config.VIP)
+	log.Info("new leader elected", "id", identity)
+}

pkg/manager/worker/worker.go

@@ -0,0 +1,60 @@
+package worker
+
+import (
+	"context"
+	"sync"
+	"sync/atomic"
+
+	"github.com/kube-vip/kube-vip/pkg/arp"
+	"github.com/kube-vip/kube-vip/pkg/bgp"
+	"github.com/kube-vip/kube-vip/pkg/election"
+	"github.com/kube-vip/kube-vip/pkg/kubevip"
+	"github.com/kube-vip/kube-vip/pkg/lease"
+	"github.com/kube-vip/kube-vip/pkg/networkinterface"
+	"github.com/kube-vip/kube-vip/pkg/node"
+	"github.com/kube-vip/kube-vip/pkg/route"
+	"github.com/kube-vip/kube-vip/pkg/services"
+	"k8s.io/client-go/kubernetes"
+)
+
+type Worker interface {
+	Configure(context.Context, *sync.WaitGroup) error
+	InitControlPlane() error
+	StartControlPlane(context.Context, *election.Manager)
+	ConfigureServices()
+	StartServices(ctx context.Context) error
+	Name() string
+	Cleanup()
+}
+
+func New(arpMgr *arp.Manager, intfMgr *networkinterface.Manager,
+	config *kubevip.Config, closing *atomic.Bool, killFunc func(),
+	svcProcessor *services.Processor, mutex *sync.Mutex, clientSet *kubernetes.Clientset,
+	bgpServer *bgp.Server, electionMgr *election.Manager, leaseMgr *lease.Manager,
+	routeMgr *route.Manager, nodeLabelMgr node.Labeler) Worker {
+	if config.EnableARP {
+		return NewARP(arpMgr, intfMgr, config, closing, killFunc,
+			svcProcessor, mutex, clientSet, electionMgr, leaseMgr, routeMgr,
+			nodeLabelMgr)
+	}
+
+	if config.EnableBGP {
+		return NewBGP(arpMgr, intfMgr, config, closing, killFunc,
+			svcProcessor, mutex, clientSet, bgpServer,
+			electionMgr, leaseMgr, routeMgr, nodeLabelMgr)
+	}
+
+	if config.EnableRoutingTable {
+		return NewTable(arpMgr, intfMgr, config, closing, killFunc,
+			svcProcessor, mutex, clientSet, electionMgr, leaseMgr, routeMgr,
+			nodeLabelMgr)
+	}
+
+	if config.EnableWireguard {
+		return NewWireGuard(arpMgr, intfMgr, config, closing, killFunc,
+			svcProcessor, mutex, clientSet, electionMgr, leaseMgr, routeMgr,
+			nodeLabelMgr)
+	}
+
+	return nil
+}

pkg/metrics/prometheus.go

@@ -0,0 +1,86 @@
+package metrics
+
+import "github.com/prometheus/client_golang/prometheus"
+
+var (
+	// Service / VIP Lifecycle
+	ActiveServices = prometheus.NewGaugeVec(
+		prometheus.GaugeOpts{Name: "kube_vip_active_services", Help: "How many LB services are currently managed"},
+		[]string{"namespace"},
+	)
+	ServiceReconcileErrorsTotal = prometheus.NewCounterVec(
+		prometheus.CounterOpts{Name: "kube_vip_service_reconcile_errors_total", Help: "Count reconcile failures"},
+		[]string{"namespace", "name", "reason"},
+	)
+	ServiceReconcileDuration = prometheus.NewHistogramVec(
+		prometheus.HistogramOpts{Name: "kube_vip_service_reconcile_duration_seconds", Help: "How long AddOrModify takes end-to-end"},
+		[]string{"namespace"},
+	)
+
+	// This is a prometheus counter used to count the number of events received
+	// from the service watcher
+	CountServiceWatchEvent = prometheus.NewCounterVec(prometheus.CounterOpts{
+		Namespace: "kube_vip",
+		Subsystem: "manager",
+		Name:      "all_services_events",
+		Help:      "Count all events fired by the service watcher categorised by event type",
+	}, []string{"type"})
+
+	// Leader Election
+	LeaderTransitionsTotal = prometheus.NewCounterVec(
+		prometheus.CounterOpts{Name: "kube_vip_leader_election_transitions_total", Help: "Frequent transitions mean instability"},
+		[]string{"lease_name"},
+	)
+	IsLeader = prometheus.NewGaugeVec(
+		prometheus.GaugeOpts{Name: "kube_vip_is_leader", Help: "1 if this node currently holds the lease"},
+		[]string{"node", "lease_name"},
+	)
+	ServiceElectionLoops = prometheus.NewGaugeVec(
+		prometheus.GaugeOpts{Name: "kube_vip_service_election_loops",
+			Help: "Live per-service leader election restart loops on this node; more than 1 per service means loops leaked"},
+		[]string{"namespace", "name"},
+	)
+	ServiceElectionAttemptsTotal = prometheus.NewCounterVec(
+		prometheus.CounterOpts{Name: "kube_vip_service_election_attempts_total",
+			Help: "Election attempts made by the per-service leader election restart loop"},
+		[]string{"namespace", "name"},
+	)
+	ServiceElectionErrorsTotal = prometheus.NewCounterVec(
+		prometheus.CounterOpts{Name: "kube_vip_service_election_errors_total",
+			Help: "Per-service leader election failures by reason"},
+		[]string{"namespace", "name", "reason"},
+	)
+
+	// This is a prometheus gauge indicating the state of the sessions.
+	// 1 means "ESTABLISHED", 0 means "NOT ESTABLISHED"
+	BGPSessionInfoGauge = prometheus.NewGaugeVec(prometheus.GaugeOpts{
+		Namespace: "kube_vip",
+		Subsystem: "manager",
+		Name:      "bgp_session_info",
+		Help:      "Display state of session by setting metric for label value with current state to 1",
+	}, []string{"state", "peer"},
+	)
+
+	// General Health
+	BuildInfo = prometheus.NewGaugeVec(
+		prometheus.GaugeOpts{Name: "kube_vip_build_info", Help: "Constant 1; track version skew across nodes"},
+		[]string{"version", "build", "node"},
+	)
+)
+
+func RegisterPrometheusMetrics() {
+	// Register all metrics with Prometheus
+	prometheus.MustRegister(
+		ActiveServices,
+		ServiceReconcileErrorsTotal,
+		ServiceReconcileDuration,
+		LeaderTransitionsTotal,
+		IsLeader,
+		ServiceElectionLoops,
+		ServiceElectionAttemptsTotal,
+		ServiceElectionErrorsTotal,
+		BGPSessionInfoGauge,
+		BuildInfo,
+		CountServiceWatchEvent,
+	)
+}

pkg/networkinterface/networkinterface.go

@@ -1,6 +1,7 @@
 package networkinterface
 
 import (
+	log "log/slog"
 	"sync"
 
 	"github.com/vishvananda/netlink"
@@ -23,6 +24,12 @@ func NewManager() *Manager {
 
 func (m *Manager) Get(intf netlink.Link) *Link {
 	if l, ok := m.interfaces[intf.Attrs().Name]; ok {
+		updated, err := netlink.LinkByName(l.Intf.Attrs().Name)
+		if err != nil {
+			log.Error("failed to get interface %q: %w", l.Intf.Attrs().Name, err)
+			return nil
+		}
+		l.Intf = updated
 		return l
 	}
 	result := &Link{

pkg/nftables/client.go

@@ -0,0 +1,379 @@
+package nftables
+
+import (
+	"bytes"
+	"errors"
+	"fmt"
+	"hash/fnv"
+	log "log/slog"
+	"net"
+	"os"
+	"reflect"
+	"slices"
+	"strings"
+
+	"github.com/google/nftables"
+	"github.com/google/nftables/expr"
+	"github.com/google/nftables/userdata"
+)
+
+const (
+	TableFilter = "filter"
+
+	ipv4SrcOffset = 12
+	ipv4DstOffset = 16
+	ipv6SrcOffset = 8
+	ipv6DstOffset = 24
+)
+
+type Client struct {
+	conn   *nftables.Conn
+	family nftables.TableFamily
+}
+
+func NewClient(family nftables.TableFamily) (*Client, error) {
+	conn, err := nftables.New()
+	if err != nil {
+		return nil, fmt.Errorf("failed to create nftables client connection: %w", err)
+	}
+	return &Client{
+		conn:   conn,
+		family: family,
+	}, nil
+}
+
+func (c *Client) Close() error {
+	if err := c.conn.CloseLasting(); err != nil {
+		return fmt.Errorf("failed to close nftables client: %w", err)
+	}
+	return nil
+}
+
+func (c *Client) Flush() {
+	c.conn.Flush()
+}
+
+func (c *Client) GetTable(name string) *nftables.Table {
+	return &nftables.Table{
+		Family: c.family,
+		Name:   name,
+	}
+}
+
+func (c *Client) GetChain(table, chain string) (*nftables.Chain, error) {
+	t := c.GetTable(table)
+	ch, err := c.conn.ListChain(t, chain)
+	if err != nil {
+		log.Error(err.Error())
+		if strings.Contains(err.Error(), "no such file") {
+			return nil, ErrChainNotFound
+		}
+		return nil, err
+	}
+	return ch, nil
+}
+
+func (c *Client) CheckChain(table, chain string) bool {
+	_, err := c.GetChain(table, chain)
+	return err == nil
+}
+
+func (c *Client) AddChain(chain *nftables.Chain) *nftables.Chain {
+	ch, err := c.GetChain(chain.Table.Name, chain.Name)
+
+	if errors.Is(err, ErrChainNotFound) {
+		log.Debug("adding chain", "name", chain.Name)
+		ch = c.conn.AddChain(chain)
+		c.Flush()
+	}
+	return ch
+}
+
+func (c *Client) DeleteChain(chain *nftables.Chain) {
+	c.conn.DelChain(chain)
+}
+
+func (c *Client) InsertUnique(rule *nftables.Rule) (*nftables.Rule, error) {
+	comment, _ := userdata.GetString(rule.UserData, userdata.TypeComment)
+	log.Debug("inserting rule", "table", rule.Table.Name, "chain", rule.Chain.Name, "comment", comment)
+	return c.add(rule, c.conn.InsertRule)
+}
+
+func (c *Client) AddUnique(rule *nftables.Rule) (*nftables.Rule, error) {
+	comment, _ := userdata.GetString(rule.UserData, userdata.TypeComment)
+	log.Debug("adding rule", "table", rule.Table.Name, "chain", rule.Chain.Name, "comment", comment)
+	return c.add(rule, c.conn.AddRule)
+}
+
+func (c *Client) add(rule *nftables.Rule, f func(r *nftables.Rule) *nftables.Rule) (*nftables.Rule, error) {
+	r, err := c.Exists(rule)
+	if err != nil {
+		return r, err
+	}
+
+	if r == nil {
+		r = f(rule)
+	}
+
+	return r, nil
+}
+
+func (c *Client) FindRuleByComment(table *nftables.Table, chain *nftables.Chain, comment string) (*nftables.Rule, error) {
+	rules, err := c.conn.GetRules(table, chain)
+	if err != nil {
+		return nil, fmt.Errorf("failed to list rules: %w", err)
+	}
+
+	ud := UserDataComment(comment)
+
+	for _, r := range rules {
+		if bytes.Equal(r.UserData, ud) {
+			return r, nil
+		}
+	}
+
+	return nil, nil
+}
+
+func (c *Client) DeleteRule(r *nftables.Rule) error {
+	if err := c.conn.DelRule(r); err != nil {
+		return fmt.Errorf("failed to delete rule: %w", err)
+	}
+	return nil
+}
+
+func (c *Client) List(t *nftables.Table, ch *nftables.Chain) ([]*nftables.Rule, error) {
+	rules, err := c.conn.GetRules(t, ch)
+	if err != nil {
+		return nil, fmt.Errorf("failed to list rules in table %q, chain %q: %w", t.Name, ch.Name, err)
+	}
+	return rules, nil
+}
+
+func (c *Client) Exists(rule *nftables.Rule) (*nftables.Rule, error) {
+	rules, err := c.conn.GetRules(rule.Table, rule.Chain)
+	if err != nil {
+		return nil, fmt.Errorf("failed to list rules in table %q, chain %q: %w", rule.Table.Name, rule.Chain.Name, err)
+	}
+
+	var existing *nftables.Rule
+	cnt := 0
+	for _, r := range rules {
+		if ruleEqual(rule, r) {
+			existing = r
+			cnt++
+		}
+	}
+
+	if cnt == 0 {
+		return nil, nil
+	}
+
+	if cnt > 1 {
+		comment, ok := userdata.GetString(rule.UserData, userdata.TypeComment)
+		if !ok {
+			log.Warn("failed to get comment for rule", "handle", rule.Handle, "table", rule.Table.Name, "chain", rule.Chain.Name)
+		} else {
+			log.Warn("too many rules", "present", cnt, "rule", comment, "table", rule.Table.Name, "chain", rule.Chain.Name)
+		}
+	}
+
+	return existing, nil
+}
+
+func (c *Client) GetLen() uint32 {
+	if c.family == nftables.TableFamilyIPv6 {
+		return net.IPv6len
+	}
+	return net.IPv4len
+}
+
+func (c *Client) GetDstOffset() uint32 {
+	if c.family == nftables.TableFamilyIPv6 {
+		return ipv6DstOffset
+	}
+	return ipv4DstOffset
+}
+
+func (c *Client) GetSrcOffset() uint32 {
+	if c.family == nftables.TableFamilyIPv6 {
+		return ipv6SrcOffset
+	}
+	return ipv4SrcOffset
+}
+
+func (c *Client) UpdateSet(set *nftables.Set, elements []nftables.SetElement) error {
+	existingSet, err := c.conn.GetSetByName(set.Table, set.Name)
+	if err != nil && !errors.Is(err, os.ErrNotExist) {
+		return fmt.Errorf("failed to get set: %w", err)
+	}
+
+	exists := err == nil && existingSet != nil
+
+	if exists {
+		existingElements, err := c.conn.GetSetElements(existingSet)
+		if err != nil {
+			return fmt.Errorf("failed to get elements for set %q, table %q", existingSet.Name, existingSet.Table.Name)
+		}
+
+		toAdd, toDel := processElements(elements, existingElements)
+
+		if len(toAdd) > 0 || len(toDel) > 0 {
+			log.Debug("updating set", "name", existingSet.Name, "table", existingSet.Table.Name)
+			if len(toDel) > 0 {
+				if err := c.conn.SetDeleteElements(existingSet, toDel); err != nil {
+					return fmt.Errorf("failed to remove elements from set %q: %w", existingSet.Name, err)
+				}
+			}
+
+			if len(toAdd) > 0 {
+				if err := c.conn.SetAddElements(existingSet, toAdd); err != nil {
+					return fmt.Errorf("failed to add elements to set %q: %w", existingSet.Name, err)
+				}
+			}
+		}
+
+		return nil
+	}
+
+	log.Debug("adding set", "name", set.Name, "table", set.Table)
+	if err := c.conn.AddSet(set, elements); err != nil {
+		return fmt.Errorf("failed to add set %q to table %q: %w", set.Name, set.Table.Name, err)
+	}
+
+	return nil
+}
+
+func (c *Client) DeleteSet(table *nftables.Table, name string) error {
+	setName, err := hash(name)
+	if err != nil {
+		return fmt.Errorf("failed to hash name: %w", err)
+	}
+
+	set, err := c.conn.GetSetByName(table, setName)
+	if err != nil {
+		return fmt.Errorf("failed to get set: %w", err)
+	}
+	if set != nil {
+		c.conn.DelSet(set)
+	}
+	return nil
+}
+
+func (c *Client) NewSet(chain *nftables.Chain, name string) (*nftables.Set, error) {
+	setName, err := hash(name)
+	if err != nil {
+		return nil, fmt.Errorf("failed to hash name: %w", err)
+	}
+
+	return &nftables.Set{
+		Table:   chain.Table,
+		Name:    setName,
+		Comment: name,
+		KeyType: nftables.TypeInetService,
+	}, nil
+}
+
+func hash(data string) (string, error) {
+	h := fnv.New32a()
+	if _, err := h.Write([]byte(data)); err != nil {
+		return "", fmt.Errorf("failed to generate hash: %w", err)
+	}
+
+	return fmt.Sprintf("%x", h.Sum32()), nil
+}
+
+func processElements(newEls, existingEls []nftables.SetElement) (toAdd, toDel []nftables.SetElement) {
+	toAdd = findNonCommon(newEls, existingEls)
+	toDel = findNonCommon(existingEls, newEls)
+	return
+}
+
+func findNonCommon(a, b []nftables.SetElement) []nftables.SetElement {
+	nonCommon := []nftables.SetElement{}
+	for i := range a {
+		if !isPresent(a[i], b) {
+			nonCommon = append(nonCommon, a[i])
+		}
+	}
+	return nonCommon
+}
+
+func isPresent(toCheck nftables.SetElement, elements []nftables.SetElement) bool {
+	for _, e := range elements {
+		if slices.Compare(toCheck.Key, e.Key) == 0 {
+			return true
+		}
+	}
+
+	return false
+}
+
+func ruleEqual(a, b *nftables.Rule) bool {
+	if a.Chain.Name != b.Chain.Name {
+		return false
+	}
+	if a.Table.Name != b.Table.Name {
+		return false
+	}
+
+	if !bytes.Equal(a.UserData, b.UserData) {
+		return false
+	}
+
+	for i := range a.Exprs {
+		switch a.Exprs[i].(type) {
+		case *expr.Meta:
+			if !exprEqual(&expr.Meta{}, a.Exprs[i], b.Exprs[i]) {
+				return false
+			}
+		case *expr.Lookup:
+			if !exprEqual(&expr.Lookup{}, a.Exprs[i], b.Exprs[i]) {
+				return false
+			}
+		case *expr.Verdict:
+			if !exprEqual(&expr.Verdict{}, a.Exprs[i], b.Exprs[i]) {
+				return false
+			}
+		case *expr.Cmp:
+			if !exprEqual(&expr.Cmp{}, a.Exprs[i], b.Exprs[i]) {
+				return false
+			}
+		case *expr.Payload:
+			if !exprEqual(&expr.Payload{}, a.Exprs[i], b.Exprs[i]) {
+				return false
+			}
+		case *expr.Ct:
+			if !exprEqual(&expr.Ct{}, a.Exprs[i], b.Exprs[i]) {
+				return false
+			}
+		case *expr.Bitwise:
+			if !exprEqual(&expr.Bitwise{}, a.Exprs[i], b.Exprs[i]) {
+				return false
+			}
+		}
+	}
+
+	return true
+}
+
+func exprEqual[V *expr.Meta | *expr.Lookup | *expr.Verdict | *expr.Cmp | *expr.Payload | *expr.Ct | *expr.Bitwise](_ V, aExpr, bExpr expr.Any) bool {
+	aExprCast, ok := aExpr.(V)
+	if !ok {
+		return false
+	}
+	bExprCast, ok := bExpr.(V)
+	if !ok {
+		return false
+	}
+	if reflect.DeepEqual(aExprCast, bExprCast) {
+		return true
+	}
+	return false
+}
+
+func UserDataComment(comment string) []byte {
+	return userdata.AppendString([]byte{}, userdata.TypeComment, comment)
+}
+
+var ErrChainNotFound = errors.New("chain not found")

pkg/nftables/nftables_test.go

@@ -0,0 +1,109 @@
+package nftables
+
+import (
+	"testing"
+
+	googlenftables "github.com/google/nftables"
+)
+
+func TestEgressTableName(t *testing.T) {
+	tests := []struct {
+		name     string
+		baseName string
+		ipv6     bool
+		want     string
+	}{
+		{name: "default IPv4", want: "kube_vip_v4"},
+		{name: "default IPv6", ipv6: true, want: "kube_vip_v6"},
+		{name: "release IPv4", baseName: "kube_vip_release_a", want: "kube_vip_release_a_v4"},
+		{name: "namespace IPv6", baseName: "kube-vip-networking", ipv6: true, want: "kube-vip-networking_v6"},
+	}
+
+	for _, tt := range tests {
+		t.Run(tt.name, func(t *testing.T) {
+			if got := egressTableName(tt.baseName, tt.ipv6); got != tt.want {
+				t.Fatalf("egressTableName(%q, %t) = %q, want %q", tt.baseName, tt.ipv6, got, tt.want)
+			}
+		})
+	}
+}
+
+func TestGetEgressTable(t *testing.T) {
+	table := GetEgressTable(false, "release_a")
+	if table.Name != "release_a_v4" {
+		t.Fatalf("table name = %q, want %q", table.Name, "release_a_v4")
+	}
+}
+
+func TestEgressTableBaseName(t *testing.T) {
+	if got := EgressTableBaseName(""); got != DefaultEgressTableName {
+		t.Fatalf("EgressTableBaseName(\"\") = %q, want %q", got, DefaultEgressTableName)
+	}
+	if got := EgressTableBaseName("release_a"); got != "release_a" {
+		t.Fatalf("EgressTableBaseName(\"release_a\") = %q, want %q", got, "release_a")
+	}
+}
+
+func TestEgressTableBaseNameForInstance(t *testing.T) {
+	if got := EgressTableBaseNameForInstance(""); got != DefaultEgressTableName {
+		t.Fatalf("EgressTableBaseNameForInstance(\"\") = %q, want %q", got, DefaultEgressTableName)
+	}
+	if got := EgressTableBaseNameForInstance("release_a"); got != "kube_vip_release_a" {
+		t.Fatalf("EgressTableBaseNameForInstance(\"release_a\") = %q, want %q", got, "kube_vip_release_a")
+	}
+}
+
+func TestEgressTableNameForInstance(t *testing.T) {
+	baseName := EgressTableBaseNameForInstance("release_a")
+	if got := egressTableName(baseName, false); got != "kube_vip_release_a_v4" {
+		t.Fatalf("IPv4 table name = %q, want %q", got, "kube_vip_release_a_v4")
+	}
+	if got := egressTableName(baseName, true); got != "kube_vip_release_a_v6" {
+		t.Fatalf("IPv6 table name = %q, want %q", got, "kube_vip_release_a_v6")
+	}
+}
+
+func TestShouldDeleteSNATChain(t *testing.T) {
+	tests := []struct {
+		name      string
+		chain     *googlenftables.Chain
+		keepTable string
+		want      bool
+	}{
+		{
+			name: "stale table for matching Service UID",
+			chain: &googlenftables.Chain{
+				Name:  "kube_vip_snat_service-a",
+				Table: &googlenftables.Table{Name: "release_a_v4"},
+			},
+			keepTable: "release_b_v4",
+			want:      true,
+		},
+		{
+			name: "current table for matching Service UID",
+			chain: &googlenftables.Chain{
+				Name:  "kube_vip_snat_service-a",
+				Table: &googlenftables.Table{Name: "release_b_v4"},
+			},
+			keepTable: "release_b_v4",
+			want:      false,
+		},
+		{
+			name: "different Service UID in stale table",
+			chain: &googlenftables.Chain{
+				Name:  "kube_vip_snat_service-b",
+				Table: &googlenftables.Table{Name: "release_a_v4"},
+			},
+			keepTable: "release_b_v4",
+			want:      false,
+		},
+	}
+
+	for _, tt := range tests {
+		t.Run(tt.name, func(t *testing.T) {
+			if got := shouldDeleteSNATChain(tt.chain, "kube_vip_snat_service-a", tt.keepTable); got != tt.want {
+				t.Fatalf("shouldDeleteSNATChain() = %t, want %t", got, tt.want)
+			}
+		})
+	}
+}

pkg/node/labeler/label_manager.go

@@ -0,0 +1,136 @@
+package labeler
+
+import (
+	"context"
+	"encoding/json"
+	"fmt"
+	"strings"
+	"time"
+
+	log "log/slog"
+
+	"github.com/kube-vip/kube-vip/pkg/kubevip"
+	"github.com/pkg/errors"
+	metav1 "k8s.io/apimachinery/pkg/apis/meta/v1"
+	"k8s.io/apimachinery/pkg/types"
+	"k8s.io/client-go/kubernetes"
+)
+
+// labelOperation is the operation to perform on the node labels
+type labelOperation string
+
+// labelOperation constants
+const (
+	labelOperationRemove labelOperation = "remove"
+	labelOperationAdd    labelOperation = "add"
+)
+
+// NewManager creates a new Label Manager for the given node
+func NewManager(nodeName string, clientSet *kubernetes.Clientset) *Manager {
+	return &Manager{
+		nodeName:  nodeName,
+		clientSet: clientSet,
+	}
+}
+
+// Manager is the label Manager for the node
+type Manager struct {
+	// nodeName is the name of the node to manage
+	nodeName string
+
+	// clientSet is the Kubernetes client set to use
+	clientSet *kubernetes.Clientset
+}
+
+// AddLabel a new label to the node
+func (m *Manager) AddLabel(labels map[string]string) error {
+	log.Debug("add labels to node", "node", m.nodeName)
+	return m.patchNode(labelOperationAdd, labels)
+}
+
+// RemoveLabel a label from the node
+func (m *Manager) RemoveLabel(labels map[string]string) error {
+	log.Debug("delete label from node")
+	return m.patchNode(labelOperationRemove, labels)
+}
+
+// CleanUpLabels purges the node labels
+func (m *Manager) CleanUpLabels(timeout time.Duration) error {
+	log.Debug("cleaning up labels for node", "node", m.nodeName, "timeout", timeout)
+
+	// create new context for labels cleanup (independent)
+	ctx, cancel := context.WithTimeout(context.Background(), timeout)
+	defer cancel()
+
+	// get the node
+	node, err := m.clientSet.CoreV1().Nodes().Get(ctx, m.nodeName, metav1.GetOptions{})
+	if err != nil {
+		return errors.Wrapf(err, "failed to get node %s", m.nodeName)
+	}
+
+	// collect all labels with the prefix to remove
+	labels := map[string]string{}
+	for _, k := range kubevip.GetKeysForCleanup() {
+		if _, ok := node.Labels[k]; ok {
+			labels[k] = ""
+		}
+	}
+
+	if len(labels) == 0 {
+		log.Debug("no labels to remove for node", "node", m.nodeName)
+		return nil
+	}
+
+	// patch the node with the labels to remove
+	return m.patchNode(labelOperationRemove, labels)
+}
+
+// patchNode patches the node with the given labels
+func (m *Manager) patchNode(operation labelOperation, labels map[string]string) error {
+	type patchStringLabel struct {
+		Op    string `json:"op"`
+		Path  string `json:"path"`
+		Value string `json:"value"`
+	}
+
+	ctx, cancel := context.WithTimeout(context.Background(), time.Second*30)
+	defer cancel()
+
+	patchLabels := []patchStringLabel{}
+	// generate the patch
+	for k, v := range labels {
+		patchLabels = append(patchLabels, patchStringLabel{
+			Op: string(operation),
+			// replace all slashes with ~1
+			Path:  fmt.Sprintf("/metadata/labels/%s", strings.ReplaceAll(k, "/", "~1")),
+			Value: v,
+		})
+	}
+
+	patchData, err := json.Marshal(patchLabels)
+	if err != nil {
+		log.Debug("node patch marshaling failed", "err", err, "labels", labels, "patch", patchLabels)
+		return errors.Wrapf(err, "node patch marshaling failed for labels %v", labels)
+	}
+
+	log.Debug("patching node",
+		"node", m.nodeName,
+		"patch", string(patchData),
+		"operation", operation,
+		"labels", labels,
+		"clientSetNil", m.clientSet == nil)
+	if m.clientSet == nil {
+		return errors.New("kubernetes client is not initialized")
+	}
+
+	// patch node
+	node, err := m.clientSet.CoreV1().Nodes().Patch(ctx, m.nodeName, types.JSONPatchType, patchData, metav1.PatchOptions{})
+	if err != nil {
+		log.Debug("node patching failed", "err", err, "patchData", patchData)
+		return errors.Wrapf(err, "node patching failed with patch %s", string(patchData))
+	}
+
+	log.Debug("updated", "node", m.nodeName, "labels", node.Labels)
+
+	return nil
+}