Aligns new tests to signature

by moving the persistence from applySync to the map state manager

.github/workflows/check-license-dependencies.yml

@@ -64,7 +64,7 @@ jobs:
           persist-credentials: false
 
       - name: Set up Go
-        uses: actions/setup-go@4a3601121dd01d1626a1e23e37211e3254c1c06c # v6.4.0
+        uses: actions/setup-go@924ae3a1cded613372ab5595356fb5720e22ba16 # v6.5.0
         with:
           go-version-file: "go.mod"
           cache: true

.github/workflows/golang-test-darwin.yml

@@ -21,13 +21,13 @@ jobs:
           persist-credentials: false
 
       - name: Install Go
-        uses: actions/setup-go@4a3601121dd01d1626a1e23e37211e3254c1c06c # v6.4.0
+        uses: actions/setup-go@924ae3a1cded613372ab5595356fb5720e22ba16 # v6.5.0
         with:
           go-version-file: "go.mod"
           cache: false
 
       - name: Cache Go modules
-        uses: actions/cache@27d5ce7f107fe9357f9df03efb73ab90386fccae # v5.0.5
+        uses: actions/cache@2c8a9bd7457de244a408f35966fab2fb45fda9c8 # v6.0.0
         with:
           path: ~/go/pkg/mod
           key: macos-gotest-${{ hashFiles('**/go.sum') }}
@@ -45,7 +45,7 @@ jobs:
         run: git --no-pager diff --exit-code
 
       - name: Test
-        run: NETBIRD_STORE_ENGINE=${{ matrix.store }} CI=true go test -coverprofile=coverage.txt -tags=devcert -exec 'sudo --preserve-env=CI,NETBIRD_STORE_ENGINE' -timeout 5m -p 1 $(go list ./... | grep -v -e /management -e /signal -e /relay -e /proxy -e /combined)
+        run: NETBIRD_STORE_ENGINE=${{ matrix.store }} CI=true go test -coverprofile=coverage.txt -tags 'devcert privileged' -exec 'sudo --preserve-env=CI,NETBIRD_STORE_ENGINE' -timeout 5m -p 1 $(go list ./... | grep -v -e /management -e /signal -e /relay -e /proxy -e /combined -e /client/testutil/privileged)
 
       - name: Upload coverage reports to Codecov
         uses: codecov/codecov-action@fb8b3582c8e4def4969c97caa2f19720cb33a72f #v7.0.0

.github/workflows/golang-test-freebsd.yml

@@ -48,14 +48,14 @@ jobs:
             export PATH=$PATH:/usr/local/go/bin:$HOME/go/bin
             time go build -o netbird client/main.go
             # check all component except management, since we do not support management server on freebsd
-            time go test -timeout 1m -failfast ./base62/...
+            time go test -tags privileged -timeout 1m -failfast ./base62/...
             # NOTE: without -p1 `client/internal/dns` will fail because of `listen udp4 :33100: bind: address already in use`
-            time go test -timeout 8m -failfast -v -p 1 ./client/...
+            time go test -tags privileged -timeout 8m -failfast -v -p 1 ./client/...
-            time go test -timeout 1m -failfast ./dns/...
+            time go test -tags privileged -timeout 1m -failfast ./dns/...
-            time go test -timeout 1m -failfast ./encryption/...
+            time go test -tags privileged -timeout 1m -failfast ./encryption/...
-            time go test -timeout 1m -failfast ./formatter/...
+            time go test -tags privileged -timeout 1m -failfast ./formatter/...
-            time go test -timeout 1m -failfast ./client/iface/...
+            time go test -tags privileged -timeout 1m -failfast ./client/iface/...
-            time go test -timeout 1m -failfast ./route/...
+            time go test -tags privileged -timeout 1m -failfast ./route/...
-            time go test -timeout 1m -failfast ./sharedsock/...
+            time go test -tags privileged -timeout 1m -failfast ./sharedsock/...
-            time go test -timeout 1m -failfast ./util/...
+            time go test -tags privileged -timeout 1m -failfast ./util/...
-            time go test -timeout 1m -failfast ./version/...
+            time go test -tags privileged -timeout 1m -failfast ./version/...

.github/workflows/golang-test-linux.yml

@@ -30,7 +30,7 @@ jobs:
               - 'management/**'
 
       - name: Install Go
-        uses: actions/setup-go@4a3601121dd01d1626a1e23e37211e3254c1c06c # v6.4.0
+        uses: actions/setup-go@924ae3a1cded613372ab5595356fb5720e22ba16 # v6.5.0
         with:
           go-version-file: "go.mod"
           cache: false
@@ -41,7 +41,7 @@ jobs:
           echo "modcache=$(go env GOMODCACHE)" >> $GITHUB_ENV
 
       - name: Cache Go modules
-        uses: actions/cache@27d5ce7f107fe9357f9df03efb73ab90386fccae # v5.0.5
+        uses: actions/cache@2c8a9bd7457de244a408f35966fab2fb45fda9c8 # v6.0.0
         id: cache
         with:
           path: |
@@ -124,7 +124,7 @@ jobs:
           persist-credentials: false
 
       - name: Install Go
-        uses: actions/setup-go@4a3601121dd01d1626a1e23e37211e3254c1c06c # v6.4.0
+        uses: actions/setup-go@924ae3a1cded613372ab5595356fb5720e22ba16 # v6.5.0
         with:
           go-version-file: "go.mod"
           cache: false
@@ -135,7 +135,7 @@ jobs:
           echo "modcache=$(go env GOMODCACHE)" >> $GITHUB_ENV
 
       - name: Cache Go modules
-        uses: actions/cache/restore@27d5ce7f107fe9357f9df03efb73ab90386fccae # v5.0.5
+        uses: actions/cache/restore@2c8a9bd7457de244a408f35966fab2fb45fda9c8 # v6.0.0
         with:
           path: |
             ${{ env.cache }}
@@ -158,7 +158,7 @@ jobs:
         run: git --no-pager diff --exit-code
 
       - name: Test
-        run: CGO_ENABLED=1 GOARCH=${{ matrix.arch }} CI=true go test -coverprofile=coverage.txt -tags devcert -exec 'sudo' -timeout 10m -p 1 $(go list ./... | grep -v -e /management -e /signal -e /relay -e /proxy -e /combined)
+        run: CGO_ENABLED=1 GOARCH=${{ matrix.arch }} CI=true go test -coverprofile=coverage.txt -tags devcert -timeout 10m -p 1 $(go list ./... | grep -v -e /management -e /signal -e /relay -e /proxy -e /combined)
 
       - name: Upload coverage reports to Codecov
         if: matrix.arch == 'amd64'
@@ -180,7 +180,7 @@ jobs:
           persist-credentials: false
 
       - name: Install Go
-        uses: actions/setup-go@4a3601121dd01d1626a1e23e37211e3254c1c06c # v6.4.0
+        uses: actions/setup-go@924ae3a1cded613372ab5595356fb5720e22ba16 # v6.5.0
         with:
           go-version-file: "go.mod"
           cache: false
@@ -192,7 +192,7 @@ jobs:
           echo "modcache_dir=$(go env GOMODCACHE)" >> $GITHUB_OUTPUT
 
       - name: Cache Go modules
-        uses: actions/cache/restore@27d5ce7f107fe9357f9df03efb73ab90386fccae # v5.0.5
+        uses: actions/cache/restore@2c8a9bd7457de244a408f35966fab2fb45fda9c8 # v6.0.0
         id: cache-restore
         with:
           path: |
@@ -229,7 +229,7 @@ jobs:
             sh -c ' \
               apk update; apk add --no-cache \
                 ca-certificates iptables ip6tables dbus dbus-dev libpcap-dev build-base; \
-              go test -buildvcs=false -tags devcert -v -timeout 10m -p 1 $(go list -buildvcs=false ./... | grep -v -e /management -e /signal -e /relay -e /proxy -e /combined -e /client/ui -e /upload-server)
+              go test -buildvcs=false -tags "devcert privileged" -v -timeout 10m -p 1 $(go list -buildvcs=false ./... | grep -v -e /management -e /signal -e /relay -e /proxy -e /combined -e /client/ui -e /upload-server -e /client/testutil/privileged)
             '
 
   test_relay:
@@ -251,7 +251,7 @@ jobs:
           persist-credentials: false
 
       - name: Install Go
-        uses: actions/setup-go@4a3601121dd01d1626a1e23e37211e3254c1c06c # v6.4.0
+        uses: actions/setup-go@924ae3a1cded613372ab5595356fb5720e22ba16 # v6.5.0
         with:
           go-version-file: "go.mod"
           cache: false
@@ -266,7 +266,7 @@ jobs:
           echo "modcache=$(go env GOMODCACHE)" >> $GITHUB_ENV
 
       - name: Cache Go modules
-        uses: actions/cache/restore@27d5ce7f107fe9357f9df03efb73ab90386fccae # v5.0.5
+        uses: actions/cache/restore@2c8a9bd7457de244a408f35966fab2fb45fda9c8 # v6.0.0
         with:
           path: |
             ${{ env.cache }}
@@ -311,7 +311,7 @@ jobs:
           persist-credentials: false
 
       - name: Install Go
-        uses: actions/setup-go@4a3601121dd01d1626a1e23e37211e3254c1c06c # v6.4.0
+        uses: actions/setup-go@924ae3a1cded613372ab5595356fb5720e22ba16 # v6.5.0
         with:
           go-version-file: "go.mod"
           cache: false
@@ -325,7 +325,7 @@ jobs:
           echo "modcache=$(go env GOMODCACHE)" >> $GITHUB_ENV
 
       - name: Cache Go modules
-        uses: actions/cache/restore@27d5ce7f107fe9357f9df03efb73ab90386fccae # v5.0.5
+        uses: actions/cache/restore@2c8a9bd7457de244a408f35966fab2fb45fda9c8 # v6.0.0
         with:
           path: |
             ${{ env.cache }}
@@ -368,7 +368,7 @@ jobs:
           persist-credentials: false
 
       - name: Install Go
-        uses: actions/setup-go@4a3601121dd01d1626a1e23e37211e3254c1c06c # v6.4.0
+        uses: actions/setup-go@924ae3a1cded613372ab5595356fb5720e22ba16 # v6.5.0
         with:
           go-version-file: "go.mod"
           cache: false
@@ -383,7 +383,7 @@ jobs:
           echo "modcache=$(go env GOMODCACHE)" >> $GITHUB_ENV
 
       - name: Cache Go modules
-        uses: actions/cache/restore@27d5ce7f107fe9357f9df03efb73ab90386fccae # v5.0.5
+        uses: actions/cache/restore@2c8a9bd7457de244a408f35966fab2fb45fda9c8 # v6.0.0
         with:
           path: |
             ${{ env.cache }}
@@ -429,7 +429,7 @@ jobs:
           persist-credentials: false
 
       - name: Install Go
-        uses: actions/setup-go@4a3601121dd01d1626a1e23e37211e3254c1c06c # v6.4.0
+        uses: actions/setup-go@924ae3a1cded613372ab5595356fb5720e22ba16 # v6.5.0
         with:
           go-version-file: "go.mod"
           cache: false
@@ -440,7 +440,7 @@ jobs:
           echo "modcache=$(go env GOMODCACHE)" >> $GITHUB_ENV
 
       - name: Cache Go modules
-        uses: actions/cache/restore@27d5ce7f107fe9357f9df03efb73ab90386fccae # v5.0.5
+        uses: actions/cache/restore@2c8a9bd7457de244a408f35966fab2fb45fda9c8 # v6.0.0
         with:
           path: |
             ${{ env.cache }}
@@ -534,7 +534,7 @@ jobs:
           persist-credentials: false
 
       - name: Install Go
-        uses: actions/setup-go@4a3601121dd01d1626a1e23e37211e3254c1c06c # v6.4.0
+        uses: actions/setup-go@924ae3a1cded613372ab5595356fb5720e22ba16 # v6.5.0
         with:
           go-version-file: "go.mod"
           cache: false
@@ -545,7 +545,7 @@ jobs:
           echo "modcache=$(go env GOMODCACHE)" >> $GITHUB_ENV
 
       - name: Cache Go modules
-        uses: actions/cache/restore@27d5ce7f107fe9357f9df03efb73ab90386fccae # v5.0.5
+        uses: actions/cache/restore@2c8a9bd7457de244a408f35966fab2fb45fda9c8 # v6.0.0
         with:
           path: |
             ${{ env.cache }}
@@ -579,10 +579,11 @@ jobs:
           CGO_ENABLED=1 GOARCH=${{ matrix.arch }} \
           NETBIRD_STORE_ENGINE=${{ matrix.store }} \
           CI=true \
-          GIT_BRANCH=${{ github.ref_name }} \
           go test -tags devcert -run=^$ -bench=. \
           -exec 'sudo --preserve-env=CI,NETBIRD_STORE_ENGINE,GIT_BRANCH,GITHUB_RUN_ID' \
           -timeout 20m ./management/... ./shared/management/... $(go list ./management/... ./shared/management/... | grep -v -e /management/server/http)
+        env:
+          GIT_BRANCH: ${{ github.ref_name }}
 
   api_benchmark:
     name: "Management / Benchmark (API)"
@@ -628,7 +629,7 @@ jobs:
           persist-credentials: false
 
       - name: Install Go
-        uses: actions/setup-go@4a3601121dd01d1626a1e23e37211e3254c1c06c # v6.4.0
+        uses: actions/setup-go@924ae3a1cded613372ab5595356fb5720e22ba16 # v6.5.0
         with:
           go-version-file: "go.mod"
           cache: false
@@ -639,7 +640,7 @@ jobs:
           echo "modcache=$(go env GOMODCACHE)" >> $GITHUB_ENV
 
       - name: Cache Go modules
-        uses: actions/cache/restore@27d5ce7f107fe9357f9df03efb73ab90386fccae # v5.0.5
+        uses: actions/cache/restore@2c8a9bd7457de244a408f35966fab2fb45fda9c8 # v6.0.0
         with:
           path: |
             ${{ env.cache }}
@@ -673,12 +674,13 @@ jobs:
           CGO_ENABLED=1 GOARCH=${{ matrix.arch }} \
           NETBIRD_STORE_ENGINE=${{ matrix.store }} \
           CI=true \
-          GIT_BRANCH=${{ github.ref_name }} \
           go test -tags=benchmark \
             -run=^$ \
             -bench=. \
             -exec 'sudo --preserve-env=CI,NETBIRD_STORE_ENGINE,GIT_BRANCH,GITHUB_RUN_ID' \
             -timeout 20m ./management/server/http/...
+        env:
+          GIT_BRANCH: ${{ github.ref_name }}
 
   api_integration_test:
     name: "Management / Integration"
@@ -697,7 +699,7 @@ jobs:
           persist-credentials: false
 
       - name: Install Go
-        uses: actions/setup-go@4a3601121dd01d1626a1e23e37211e3254c1c06c # v6.4.0
+        uses: actions/setup-go@924ae3a1cded613372ab5595356fb5720e22ba16 # v6.5.0
         with:
           go-version-file: "go.mod"
           cache: false
@@ -708,7 +710,7 @@ jobs:
           echo "modcache=$(go env GOMODCACHE)" >> $GITHUB_ENV
 
       - name: Cache Go modules
-        uses: actions/cache/restore@27d5ce7f107fe9357f9df03efb73ab90386fccae # v5.0.5
+        uses: actions/cache/restore@2c8a9bd7457de244a408f35966fab2fb45fda9c8 # v6.0.0
         with:
           path: |
             ${{ env.cache }}

.github/workflows/golang-test-windows.yml

@@ -23,7 +23,7 @@ jobs:
           persist-credentials: false
 
       - name: Install Go
-        uses: actions/setup-go@4a3601121dd01d1626a1e23e37211e3254c1c06c # v6.4.0
+        uses: actions/setup-go@924ae3a1cded613372ab5595356fb5720e22ba16 # v6.5.0
         id: go
         with:
           go-version-file: "go.mod"
@@ -35,7 +35,7 @@ jobs:
           echo "modcache=$(go env GOMODCACHE)" >> $env:GITHUB_ENV
 
       - name: Cache Go modules
-        uses: actions/cache@27d5ce7f107fe9357f9df03efb73ab90386fccae # v5.0.5
+        uses: actions/cache@2c8a9bd7457de244a408f35966fab2fb45fda9c8 # v6.0.0
         with:
           path: |
             ${{ env.cache }}
@@ -68,7 +68,7 @@ jobs:
         run: |
           $packages = go list ./... | Where-Object { $_ -notmatch '/management' } | Where-Object { $_ -notmatch '/relay' } | Where-Object { $_ -notmatch '/signal' } | Where-Object { $_ -notmatch '/proxy' } | Where-Object { $_ -notmatch '/combined' }
           $goExe = "C:\hostedtoolcache\windows\go\${{ steps.go.outputs.go-version }}\x64\bin\go.exe"
-          $cmd = "$goExe test -tags=devcert -timeout 10m -p 1 $($packages -join ' ') > test-out.txt 2>&1"
+          $cmd = "$goExe test -tags `"devcert privileged`" -timeout 10m -p 1 $($packages -join ' ') > test-out.txt 2>&1"
           Set-Content -Path "${{ github.workspace }}\run-tests.cmd" -Value $cmd
 
       - name: test

.github/workflows/golangci-lint.yml

@@ -48,7 +48,7 @@ jobs:
         run: |
           ! awk '/const \(/,/)/{print $0}' management/server/activity/codes.go | grep -o '= [0-9]*' | sort | uniq -d | grep .
       - name: Install Go
-        uses: actions/setup-go@4a3601121dd01d1626a1e23e37211e3254c1c06c # v6.4.0
+        uses: actions/setup-go@924ae3a1cded613372ab5595356fb5720e22ba16 # v6.5.0
         with:
           go-version-file: "go.mod"
           cache: false

.github/workflows/mobile-build-validation.yml

@@ -20,7 +20,7 @@ jobs:
         with:
           persist-credentials: false
       - name: Install Go
-        uses: actions/setup-go@4a3601121dd01d1626a1e23e37211e3254c1c06c # v6.4.0
+        uses: actions/setup-go@924ae3a1cded613372ab5595356fb5720e22ba16 # v6.5.0
         with:
           go-version-file: "go.mod"
       - name: Setup Android SDK
@@ -28,13 +28,13 @@ jobs:
         with:
           cmdline-tools-version: 8512546
       - name: Setup Java
-        uses: actions/setup-java@ad2b38190b15e4d6bdf0c97fb4fca8412226d287
+        uses: actions/setup-java@1bcf9fb12cf4aa7d266a90ae39939e61372fe520
         with:
           java-version: "11"
           distribution: "adopt"
       - name: NDK Cache
         id: ndk-cache
-        uses: actions/cache@27d5ce7f107fe9357f9df03efb73ab90386fccae # v5.0.5
+        uses: actions/cache@2c8a9bd7457de244a408f35966fab2fb45fda9c8 # v6.0.0
         with:
           path: /usr/local/lib/android/sdk/ndk
           key: ndk-cache-23.1.7779620
@@ -58,7 +58,7 @@ jobs:
         with:
           persist-credentials: false
       - name: Install Go
-        uses: actions/setup-go@4a3601121dd01d1626a1e23e37211e3254c1c06c # v6.4.0
+        uses: actions/setup-go@924ae3a1cded613372ab5595356fb5720e22ba16 # v6.5.0
         with:
           go-version-file: "go.mod"
       - name: install gomobile

.github/workflows/release.yml

@@ -166,12 +166,12 @@ jobs:
           fi
 
       - name: Set up Go
-        uses: actions/setup-go@4a3601121dd01d1626a1e23e37211e3254c1c06c # v6.4.0
+        uses: actions/setup-go@924ae3a1cded613372ab5595356fb5720e22ba16 # v6.5.0
         with:
           go-version-file: "go.mod"
           cache: false
       - name: Cache Go modules
-        uses: actions/cache/restore@27d5ce7f107fe9357f9df03efb73ab90386fccae # v5.0.5
+        uses: actions/cache/restore@2c8a9bd7457de244a408f35966fab2fb45fda9c8 # v6.0.0
         with:
           path: |
             ~/go/pkg/mod
@@ -374,12 +374,12 @@ jobs:
           fi
 
       - name: Set up Go
-        uses: actions/setup-go@4a3601121dd01d1626a1e23e37211e3254c1c06c # v6.4.0
+        uses: actions/setup-go@924ae3a1cded613372ab5595356fb5720e22ba16 # v6.5.0
         with:
           go-version-file: "go.mod"
           cache: false
       - name: Cache Go modules
-        uses: actions/cache@27d5ce7f107fe9357f9df03efb73ab90386fccae # v5.0.5
+        uses: actions/cache@2c8a9bd7457de244a408f35966fab2fb45fda9c8 # v6.0.0
         with:
           path: |
             ~/go/pkg/mod
@@ -469,12 +469,12 @@ jobs:
           fetch-depth: 0 # It is required for GoReleaser to work properly
           persist-credentials: false
       - name: Set up Go
-        uses: actions/setup-go@4a3601121dd01d1626a1e23e37211e3254c1c06c # v6.4.0
+        uses: actions/setup-go@924ae3a1cded613372ab5595356fb5720e22ba16 # v6.5.0
         with:
           go-version-file: "go.mod"
           cache: false
       - name: Cache Go modules
-        uses: actions/cache@27d5ce7f107fe9357f9df03efb73ab90386fccae # v5.0.5
+        uses: actions/cache@2c8a9bd7457de244a408f35966fab2fb45fda9c8 # v6.0.0
         with:
           path: |
             ~/go/pkg/mod

.github/workflows/test-infrastructure-files.yml

@@ -73,12 +73,12 @@ jobs:
           persist-credentials: false
 
       - name: Install Go
-        uses: actions/setup-go@4a3601121dd01d1626a1e23e37211e3254c1c06c # v6.4.0
+        uses: actions/setup-go@924ae3a1cded613372ab5595356fb5720e22ba16 # v6.5.0
         with:
           go-version-file: "go.mod"
 
       - name: Cache Go modules
-        uses: actions/cache@27d5ce7f107fe9357f9df03efb73ab90386fccae # v5.0.5
+        uses: actions/cache@2c8a9bd7457de244a408f35966fab2fb45fda9c8 # v6.0.0
         with:
           path: ~/go/pkg/mod
           key: ${{ runner.os }}-go-${{ hashFiles('**/go.sum') }}

.github/workflows/wasm-build-validation.yml

@@ -23,7 +23,7 @@ jobs:
         with:
           persist-credentials: false
       - name: Install Go
-        uses: actions/setup-go@4a3601121dd01d1626a1e23e37211e3254c1c06c # v6.4.0
+        uses: actions/setup-go@924ae3a1cded613372ab5595356fb5720e22ba16 # v6.5.0
         with:
           go-version-file: "go.mod"
       - name: Install dependencies
@@ -48,7 +48,7 @@ jobs:
         with:
           persist-credentials: false
       - name: Install Go
-        uses: actions/setup-go@4a3601121dd01d1626a1e23e37211e3254c1c06c # v6.4.0
+        uses: actions/setup-go@924ae3a1cded613372ab5595356fb5720e22ba16 # v6.5.0
         with:
           go-version-file: "go.mod"
       - name: Build Wasm client

Makefile

@@ -1,4 +1,4 @@
-.PHONY: lint lint-all lint-install setup-hooks
+.PHONY: lint lint-all lint-install setup-hooks test-unit test-privileged
 GOLANGCI_LINT := $(shell pwd)/bin/golangci-lint
 
 # Install golangci-lint locally if needed
@@ -25,3 +25,15 @@ setup-hooks:
 	@git config core.hooksPath .githooks
 	@chmod +x .githooks/pre-push
 	@echo "✅ Git hooks configured! Pre-push will now run 'make lint'"
+
+# Host-safe unit tests: excludes the privileged-tagged tests (root / system-mutating).
+# Runs as a normal user with no sudo and leaves host networking untouched.
+test-unit:
+	@go test -tags devcert -timeout 10m ./...
+
+# Privileged suite: runs the `privileged`-tagged tests inside a --privileged
+# --cap-add=NET_ADMIN container via the ory/dockertest harness. Requires Docker.
+# Narrow the run with env vars, e.g.:
+#   PRIV_RUN=TestNftablesManager PRIV_PKGS=./client/firewall/nftables/... make test-privileged
+test-privileged:
+	@go test -tags 'devcert privileged' -timeout 30m -run TestRunPrivilegedSuiteInDocker -v ./client/testutil/privileged/...

README.md

@@ -37,6 +37,11 @@
   </strong>
 </p>
 
+> ### 🤖 NetBird Agent Network (Beta)
+> Identity-aware access control for AI agents — keyless access to LLM APIs and private
+> resources over the encrypted NetBird tunnel. See [`agent-network/`](agent-network/) or
+> read the docs at **[netbird.ai](https://netbird.ai)**.
+
 **NetBird combines a configuration-free peer-to-peer private network and a centralized access control system in a single platform, making it easy to create secure private networks for your organization or home.**
 
 **Connect.** NetBird creates a WireGuard-based overlay network that automatically connects your machines over an encrypted tunnel, leaving behind the hassle of opening ports, complex firewall rules, VPN gateways, and so forth.

agent-network/README.md

@@ -0,0 +1,39 @@
+# NetBird Agent Network
+
+Agent Network is NetBird's access control layer for AI agents and the people who run
+them. It gives every agent a real identity, tied to your identity provider (IdP), and
+governs what it can reach — the LLM APIs and AI gateways it can call, and the internal
+resources it can access. Traffic flows only over the encrypted NetBird tunnel, scoped by
+policy, with no API keys to leak.
+
+> **Beta.** Agent Network is open source and can be self-hosted on your own
+> infrastructure.
+
+## How it works
+
+Agent Network is built on two existing NetBird capabilities:
+
+- **Overlay network** — the encrypted WireGuard mesh between peers.
+- **Reverse proxy** — a NetBird peer that terminates LLM requests, establishes the
+  caller's identity, evaluates policies/limits/guardrails, injects the upstream provider
+  key server-side, forwards to the API or gateway, and records usage.
+
+LLM traffic is routed through the proxy's identity-aware pipeline, while internal
+resources (databases, internal APIs, self-hosted models) are reached directly over
+peer-to-peer WireGuard tunnels, governed by the same identities and access policies.
+
+## Where the code lives
+
+There is no separate "agent-network" service — it reuses the reverse-proxy and management
+components:
+
+- [`proxy/`](../proxy) — the NetBird reverse proxy that serves the agent network endpoint
+  and runs the per-request middleware pipeline.
+- [`management/internals/modules/reverseproxy/`](../management/internals/modules/reverseproxy)
+  — the management-side control plane: providers, policies, guardrails, limits, routing,
+  and usage/access logs.
+
+## Documentation
+
+Full documentation, architecture, and quickstart:
+**https://docs.netbird.io/agent-network**

client/cmd/service_privileged_test.go

@@ -0,0 +1,196 @@
+//go:build privileged
+
+package cmd
+
+import (
+	"context"
+	"fmt"
+	"os"
+	"runtime"
+	"testing"
+	"time"
+
+	"github.com/kardianos/service"
+	"github.com/stretchr/testify/assert"
+	"github.com/stretchr/testify/require"
+)
+
+const (
+	serviceStartTimeout = 10 * time.Second
+	serviceStopTimeout  = 5 * time.Second
+	statusPollInterval  = 500 * time.Millisecond
+)
+
+// waitForServiceStatus waits for service to reach expected status with timeout
+func waitForServiceStatus(expectedStatus service.Status, timeout time.Duration) (bool, error) {
+	cfg, err := newSVCConfig()
+	if err != nil {
+		return false, err
+	}
+
+	ctxSvc, cancel := context.WithCancel(context.Background())
+	defer cancel()
+
+	s, err := newSVC(newProgram(ctxSvc, cancel), cfg)
+	if err != nil {
+		return false, err
+	}
+
+	ctx, timeoutCancel := context.WithTimeout(context.Background(), timeout)
+	defer timeoutCancel()
+
+	ticker := time.NewTicker(statusPollInterval)
+	defer ticker.Stop()
+
+	for {
+		select {
+		case <-ctx.Done():
+			return false, fmt.Errorf("timeout waiting for service status %v", expectedStatus)
+		case <-ticker.C:
+			status, err := s.Status()
+			if err != nil {
+				// Continue polling on transient errors
+				continue
+			}
+			if status == expectedStatus {
+				return true, nil
+			}
+		}
+	}
+}
+
+// TestServiceLifecycle tests the complete service lifecycle
+func TestServiceLifecycle(t *testing.T) {
+	// TODO: Add support for Windows and macOS
+	if runtime.GOOS != "linux" && runtime.GOOS != "freebsd" {
+		t.Skipf("Skipping service lifecycle test on unsupported OS: %s", runtime.GOOS)
+	}
+
+	if os.Getenv("CONTAINER") == "true" {
+		t.Skip("Skipping service lifecycle test in container environment")
+	}
+
+	originalServiceName := serviceName
+	serviceName = "netbirdtest" + fmt.Sprintf("%d", time.Now().Unix())
+	defer func() {
+		serviceName = originalServiceName
+	}()
+
+	tempDir := t.TempDir()
+	configPath = fmt.Sprintf("%s/netbird-test-config.json", tempDir)
+	logLevel = "info"
+	daemonAddr = fmt.Sprintf("unix://%s/netbird-test.sock", tempDir)
+
+	// Ensure cleanup even if a subtest fails and Stop/Uninstall subtests don't run.
+	t.Cleanup(func() {
+		cfg, err := newSVCConfig()
+		if err != nil {
+			t.Errorf("cleanup: create service config: %v", err)
+			return
+		}
+		ctxSvc, cancel := context.WithCancel(context.Background())
+		defer cancel()
+		s, err := newSVC(newProgram(ctxSvc, cancel), cfg)
+		if err != nil {
+			t.Errorf("cleanup: create service: %v", err)
+			return
+		}
+
+		// If the subtests already cleaned up, there's nothing to do.
+		if _, err := s.Status(); err != nil {
+			return
+		}
+
+		if err := s.Stop(); err != nil {
+			t.Errorf("cleanup: stop service: %v", err)
+		}
+		if err := s.Uninstall(); err != nil {
+			t.Errorf("cleanup: uninstall service: %v", err)
+		}
+	})
+
+	ctx := context.Background()
+
+	t.Run("Install", func(t *testing.T) {
+		installCmd.SetContext(ctx)
+		err := installCmd.RunE(installCmd, []string{})
+		require.NoError(t, err)
+
+		cfg, err := newSVCConfig()
+		require.NoError(t, err)
+
+		ctxSvc, cancel := context.WithCancel(context.Background())
+		defer cancel()
+
+		s, err := newSVC(newProgram(ctxSvc, cancel), cfg)
+		require.NoError(t, err)
+
+		status, err := s.Status()
+		assert.NoError(t, err)
+		assert.NotEqual(t, service.StatusUnknown, status)
+	})
+
+	t.Run("Start", func(t *testing.T) {
+		startCmd.SetContext(ctx)
+		err := startCmd.RunE(startCmd, []string{})
+		require.NoError(t, err)
+
+		running, err := waitForServiceStatus(service.StatusRunning, serviceStartTimeout)
+		require.NoError(t, err)
+		assert.True(t, running)
+	})
+
+	t.Run("Restart", func(t *testing.T) {
+		restartCmd.SetContext(ctx)
+		err := restartCmd.RunE(restartCmd, []string{})
+		require.NoError(t, err)
+
+		running, err := waitForServiceStatus(service.StatusRunning, serviceStartTimeout)
+		require.NoError(t, err)
+		assert.True(t, running)
+	})
+
+	t.Run("Reconfigure", func(t *testing.T) {
+		originalLogLevel := logLevel
+		logLevel = "debug"
+		defer func() {
+			logLevel = originalLogLevel
+		}()
+
+		reconfigureCmd.SetContext(ctx)
+		err := reconfigureCmd.RunE(reconfigureCmd, []string{})
+		require.NoError(t, err)
+
+		running, err := waitForServiceStatus(service.StatusRunning, serviceStartTimeout)
+		require.NoError(t, err)
+		assert.True(t, running)
+	})
+
+	t.Run("Stop", func(t *testing.T) {
+		stopCmd.SetContext(ctx)
+		err := stopCmd.RunE(stopCmd, []string{})
+		require.NoError(t, err)
+
+		stopped, err := waitForServiceStatus(service.StatusStopped, serviceStopTimeout)
+		require.NoError(t, err)
+		assert.True(t, stopped)
+	})
+
+	t.Run("Uninstall", func(t *testing.T) {
+		uninstallCmd.SetContext(ctx)
+		err := uninstallCmd.RunE(uninstallCmd, []string{})
+		require.NoError(t, err)
+
+		cfg, err := newSVCConfig()
+		require.NoError(t, err)
+
+		ctxSvc, cancel := context.WithCancel(context.Background())
+		defer cancel()
+
+		s, err := newSVC(newProgram(ctxSvc, cancel), cfg)
+		require.NoError(t, err)
+
+		_, err = s.Status()
+		assert.Error(t, err)
+	})
+}

client/cmd/service_test.go

@@ -1,16 +1,12 @@
 package cmd
 
 import (
-	"context"
-	"fmt"
 	"os"
 	"os/signal"
 	"runtime"
 	"syscall"
 	"testing"
-	"time"
 
-	"github.com/kardianos/service"
 	"github.com/stretchr/testify/assert"
 	"github.com/stretchr/testify/require"
 )
@@ -31,186 +27,6 @@ func TestMain(m *testing.M) {
 	os.Exit(m.Run())
 }
 
-const (
-	serviceStartTimeout = 10 * time.Second
-	serviceStopTimeout  = 5 * time.Second
-	statusPollInterval  = 500 * time.Millisecond
-)
-
-// waitForServiceStatus waits for service to reach expected status with timeout
-func waitForServiceStatus(expectedStatus service.Status, timeout time.Duration) (bool, error) {
-	cfg, err := newSVCConfig()
-	if err != nil {
-		return false, err
-	}
-
-	ctxSvc, cancel := context.WithCancel(context.Background())
-	defer cancel()
-
-	s, err := newSVC(newProgram(ctxSvc, cancel), cfg)
-	if err != nil {
-		return false, err
-	}
-
-	ctx, timeoutCancel := context.WithTimeout(context.Background(), timeout)
-	defer timeoutCancel()
-
-	ticker := time.NewTicker(statusPollInterval)
-	defer ticker.Stop()
-
-	for {
-		select {
-		case <-ctx.Done():
-			return false, fmt.Errorf("timeout waiting for service status %v", expectedStatus)
-		case <-ticker.C:
-			status, err := s.Status()
-			if err != nil {
-				// Continue polling on transient errors
-				continue
-			}
-			if status == expectedStatus {
-				return true, nil
-			}
-		}
-	}
-}
-
-// TestServiceLifecycle tests the complete service lifecycle
-func TestServiceLifecycle(t *testing.T) {
-	// TODO: Add support for Windows and macOS
-	if runtime.GOOS != "linux" && runtime.GOOS != "freebsd" {
-		t.Skipf("Skipping service lifecycle test on unsupported OS: %s", runtime.GOOS)
-	}
-
-	if os.Getenv("CONTAINER") == "true" {
-		t.Skip("Skipping service lifecycle test in container environment")
-	}
-
-	originalServiceName := serviceName
-	serviceName = "netbirdtest" + fmt.Sprintf("%d", time.Now().Unix())
-	defer func() {
-		serviceName = originalServiceName
-	}()
-
-	tempDir := t.TempDir()
-	configPath = fmt.Sprintf("%s/netbird-test-config.json", tempDir)
-	logLevel = "info"
-	daemonAddr = fmt.Sprintf("unix://%s/netbird-test.sock", tempDir)
-
-	// Ensure cleanup even if a subtest fails and Stop/Uninstall subtests don't run.
-	t.Cleanup(func() {
-		cfg, err := newSVCConfig()
-		if err != nil {
-			t.Errorf("cleanup: create service config: %v", err)
-			return
-		}
-		ctxSvc, cancel := context.WithCancel(context.Background())
-		defer cancel()
-		s, err := newSVC(newProgram(ctxSvc, cancel), cfg)
-		if err != nil {
-			t.Errorf("cleanup: create service: %v", err)
-			return
-		}
-
-		// If the subtests already cleaned up, there's nothing to do.
-		if _, err := s.Status(); err != nil {
-			return
-		}
-
-		if err := s.Stop(); err != nil {
-			t.Errorf("cleanup: stop service: %v", err)
-		}
-		if err := s.Uninstall(); err != nil {
-			t.Errorf("cleanup: uninstall service: %v", err)
-		}
-	})
-
-	ctx := context.Background()
-
-	t.Run("Install", func(t *testing.T) {
-		installCmd.SetContext(ctx)
-		err := installCmd.RunE(installCmd, []string{})
-		require.NoError(t, err)
-
-		cfg, err := newSVCConfig()
-		require.NoError(t, err)
-
-		ctxSvc, cancel := context.WithCancel(context.Background())
-		defer cancel()
-
-		s, err := newSVC(newProgram(ctxSvc, cancel), cfg)
-		require.NoError(t, err)
-
-		status, err := s.Status()
-		assert.NoError(t, err)
-		assert.NotEqual(t, service.StatusUnknown, status)
-	})
-
-	t.Run("Start", func(t *testing.T) {
-		startCmd.SetContext(ctx)
-		err := startCmd.RunE(startCmd, []string{})
-		require.NoError(t, err)
-
-		running, err := waitForServiceStatus(service.StatusRunning, serviceStartTimeout)
-		require.NoError(t, err)
-		assert.True(t, running)
-	})
-
-	t.Run("Restart", func(t *testing.T) {
-		restartCmd.SetContext(ctx)
-		err := restartCmd.RunE(restartCmd, []string{})
-		require.NoError(t, err)
-
-		running, err := waitForServiceStatus(service.StatusRunning, serviceStartTimeout)
-		require.NoError(t, err)
-		assert.True(t, running)
-	})
-
-	t.Run("Reconfigure", func(t *testing.T) {
-		originalLogLevel := logLevel
-		logLevel = "debug"
-		defer func() {
-			logLevel = originalLogLevel
-		}()
-
-		reconfigureCmd.SetContext(ctx)
-		err := reconfigureCmd.RunE(reconfigureCmd, []string{})
-		require.NoError(t, err)
-
-		running, err := waitForServiceStatus(service.StatusRunning, serviceStartTimeout)
-		require.NoError(t, err)
-		assert.True(t, running)
-	})
-
-	t.Run("Stop", func(t *testing.T) {
-		stopCmd.SetContext(ctx)
-		err := stopCmd.RunE(stopCmd, []string{})
-		require.NoError(t, err)
-
-		stopped, err := waitForServiceStatus(service.StatusStopped, serviceStopTimeout)
-		require.NoError(t, err)
-		assert.True(t, stopped)
-	})
-
-	t.Run("Uninstall", func(t *testing.T) {
-		uninstallCmd.SetContext(ctx)
-		err := uninstallCmd.RunE(uninstallCmd, []string{})
-		require.NoError(t, err)
-
-		cfg, err := newSVCConfig()
-		require.NoError(t, err)
-
-		ctxSvc, cancel := context.WithCancel(context.Background())
-		defer cancel()
-
-		s, err := newSVC(newProgram(ctxSvc, cancel), cfg)
-		require.NoError(t, err)
-
-		_, err = s.Status()
-		assert.Error(t, err)
-	})
-}
-
 // TestServiceEnvVars tests environment variable parsing
 func TestServiceEnvVars(t *testing.T) {
 	tests := []struct {

client/firewall/iptables/manager_linux_test.go

@@ -1,3 +1,5 @@
+//go:build privileged
+
 package iptables
 
 import (

client/firewall/iptables/router_linux_test.go

@@ -1,4 +1,4 @@
-//go:build !android
+//go:build !android && privileged
 
 package iptables
 

client/firewall/nftables/manager_linux_test.go

@@ -1,3 +1,5 @@
+//go:build privileged
+
 package nftables
 
 import (

client/firewall/nftables/router_linux_test.go

@@ -1,4 +1,4 @@
-//go:build !android
+//go:build !android && privileged
 
 package nftables
 

client/iface/iface_test.go

@@ -1,3 +1,5 @@
+//go:build privileged
+
 package iface
 
 import (

client/iface/wgproxy/proxy_linux_test.go

@@ -1,4 +1,4 @@
-//go:build linux && !android
+//go:build linux && !android && privileged
 
 package wgproxy
 

client/iface/wgproxy/proxy_seed_test.go

@@ -1,4 +1,4 @@
-//go:build !linux
+//go:build !linux || !privileged
 
 package wgproxy
 

client/iface/wgproxy/redirect_test.go

@@ -1,4 +1,4 @@
-//go:build linux && !android
+//go:build linux && !android && privileged
 
 package wgproxy
 
@@ -26,64 +26,6 @@ func compareUDPAddr(addr1, addr2 net.Addr) bool {
 	return udpAddr1.IP.Equal(udpAddr2.IP) && udpAddr1.Port == udpAddr2.Port
 }
 
-// TestRedirectAs_eBPF_IPv4 tests RedirectAs with eBPF proxy using IPv4 addresses
-func TestRedirectAs_eBPF_IPv4(t *testing.T) {
-	wgPort := 51850
-	ebpfProxy := ebpf.NewWGEBPFProxy(wgPort, 1280)
-	if err := ebpfProxy.Listen(); err != nil {
-		t.Fatalf("failed to initialize ebpf proxy: %v", err)
-	}
-	defer func() {
-		if err := ebpfProxy.Free(); err != nil {
-			t.Errorf("failed to free ebpf proxy: %v", err)
-		}
-	}()
-
-	proxy := ebpf.NewProxyWrapper(ebpfProxy)
-
-	// NetBird UDP address of the remote peer
-	nbAddr := &net.UDPAddr{
-		IP:   net.ParseIP("100.108.111.177"),
-		Port: 38746,
-	}
-
-	p2pEndpoint := &net.UDPAddr{
-		IP:   net.ParseIP("192.168.0.56"),
-		Port: 51820,
-	}
-
-	testRedirectAs(t, proxy, wgPort, nbAddr, p2pEndpoint)
-}
-
-// TestRedirectAs_eBPF_IPv6 tests RedirectAs with eBPF proxy using IPv6 addresses
-func TestRedirectAs_eBPF_IPv6(t *testing.T) {
-	wgPort := 51851
-	ebpfProxy := ebpf.NewWGEBPFProxy(wgPort, 1280)
-	if err := ebpfProxy.Listen(); err != nil {
-		t.Fatalf("failed to initialize ebpf proxy: %v", err)
-	}
-	defer func() {
-		if err := ebpfProxy.Free(); err != nil {
-			t.Errorf("failed to free ebpf proxy: %v", err)
-		}
-	}()
-
-	proxy := ebpf.NewProxyWrapper(ebpfProxy)
-
-	// NetBird UDP address of the remote peer
-	nbAddr := &net.UDPAddr{
-		IP:   net.ParseIP("100.108.111.177"),
-		Port: 38746,
-	}
-
-	p2pEndpoint := &net.UDPAddr{
-		IP:   net.ParseIP("fe80::56"),
-		Port: 51820,
-	}
-
-	testRedirectAs(t, proxy, wgPort, nbAddr, p2pEndpoint)
-}
-
 // TestRedirectAs_UDP_IPv4 tests RedirectAs with UDP proxy using IPv4 addresses
 func TestRedirectAs_UDP_IPv4(t *testing.T) {
 	wgPort := 51852
@@ -256,6 +198,64 @@ func testRedirectAs(t *testing.T, proxy Proxy, wgPort int, nbAddr, p2pEndpoint *
 	}
 }
 
+// TestRedirectAs_eBPF_IPv4 tests RedirectAs with eBPF proxy using IPv4 addresses
+func TestRedirectAs_eBPF_IPv4(t *testing.T) {
+	wgPort := 51850
+	ebpfProxy := ebpf.NewWGEBPFProxy(wgPort, 1280)
+	if err := ebpfProxy.Listen(); err != nil {
+		t.Fatalf("failed to initialize ebpf proxy: %v", err)
+	}
+	defer func() {
+		if err := ebpfProxy.Free(); err != nil {
+			t.Errorf("failed to free ebpf proxy: %v", err)
+		}
+	}()
+
+	proxy := ebpf.NewProxyWrapper(ebpfProxy)
+
+	// NetBird UDP address of the remote peer
+	nbAddr := &net.UDPAddr{
+		IP:   net.ParseIP("100.108.111.177"),
+		Port: 38746,
+	}
+
+	p2pEndpoint := &net.UDPAddr{
+		IP:   net.ParseIP("192.168.0.56"),
+		Port: 51820,
+	}
+
+	testRedirectAs(t, proxy, wgPort, nbAddr, p2pEndpoint)
+}
+
+// TestRedirectAs_eBPF_IPv6 tests RedirectAs with eBPF proxy using IPv6 addresses
+func TestRedirectAs_eBPF_IPv6(t *testing.T) {
+	wgPort := 51851
+	ebpfProxy := ebpf.NewWGEBPFProxy(wgPort, 1280)
+	if err := ebpfProxy.Listen(); err != nil {
+		t.Fatalf("failed to initialize ebpf proxy: %v", err)
+	}
+	defer func() {
+		if err := ebpfProxy.Free(); err != nil {
+			t.Errorf("failed to free ebpf proxy: %v", err)
+		}
+	}()
+
+	proxy := ebpf.NewProxyWrapper(ebpfProxy)
+
+	// NetBird UDP address of the remote peer
+	nbAddr := &net.UDPAddr{
+		IP:   net.ParseIP("100.108.111.177"),
+		Port: 38746,
+	}
+
+	p2pEndpoint := &net.UDPAddr{
+		IP:   net.ParseIP("fe80::56"),
+		Port: 51820,
+	}
+
+	testRedirectAs(t, proxy, wgPort, nbAddr, p2pEndpoint)
+}
+
 // TestRedirectAs_Multiple_Switches tests switching between multiple endpoints
 func TestRedirectAs_Multiple_Switches(t *testing.T) {
 	wgPort := 51856

client/internal/dns/server_privileged_test.go

@@ -0,0 +1,485 @@
+//go:build privileged
+
+package dns
+
+import (
+	"context"
+	"fmt"
+	"net/netip"
+	"os"
+	"testing"
+
+	"github.com/golang/mock/gomock"
+	"github.com/miekg/dns"
+	"github.com/stretchr/testify/assert"
+	"golang.zx2c4.com/wireguard/wgctrl/wgtypes"
+
+	"github.com/netbirdio/netbird/client/iface"
+	pfmock "github.com/netbirdio/netbird/client/iface/mocks"
+	"github.com/netbirdio/netbird/client/iface/wgaddr"
+	"github.com/netbirdio/netbird/client/internal/dns/local"
+	"github.com/netbirdio/netbird/client/internal/dns/test"
+	"github.com/netbirdio/netbird/client/internal/peer"
+	"github.com/netbirdio/netbird/client/internal/stdnet"
+	nbdns "github.com/netbirdio/netbird/dns"
+)
+
+func TestUpdateDNSServer(t *testing.T) {
+
+	nameServers := []nbdns.NameServer{
+		{
+			IP:     netip.MustParseAddr("8.8.8.8"),
+			NSType: nbdns.UDPNameServerType,
+			Port:   53,
+		},
+		{
+			IP:     netip.MustParseAddr("8.8.4.4"),
+			NSType: nbdns.UDPNameServerType,
+			Port:   53,
+		},
+	}
+
+	testCases := []struct {
+		name                string
+		initUpstreamMap     []handlerWrapper
+		initLocalZones      []nbdns.CustomZone
+		initSerial          uint64
+		inputSerial         uint64
+		inputUpdate         nbdns.Config
+		shouldFail          bool
+		expectedUpstreamMap []handlerWrapper
+		expectedLocalQs     []dns.Question
+	}{
+		{
+			name:            "Initial Config Should Succeed",
+			initUpstreamMap: nil,
+			initSerial:      0,
+			inputSerial:     1,
+			inputUpdate: nbdns.Config{
+				ServiceEnable: true,
+				CustomZones: []nbdns.CustomZone{
+					{
+						Domain:  "netbird.cloud",
+						Records: zoneRecords,
+					},
+				},
+				NameServerGroups: []*nbdns.NameServerGroup{
+					{
+						Domains:     []string{"netbird.io"},
+						NameServers: nameServers,
+					},
+					{
+						NameServers: nameServers,
+						Primary:     true,
+					},
+				},
+			},
+			expectedUpstreamMap: []handlerWrapper{
+				{
+					domain:   "netbird.io",
+					priority: PriorityUpstream,
+				},
+				{
+					domain:   "netbird.cloud",
+					priority: PriorityLocal,
+				},
+				{
+					domain:   nbdns.RootZone,
+					priority: PriorityDefault,
+				},
+			},
+			expectedLocalQs: []dns.Question{{Name: "peera.netbird.cloud.", Qtype: dns.TypeA, Qclass: dns.ClassINET}},
+		},
+		{
+			name:           "New Config Should Succeed",
+			initLocalZones: []nbdns.CustomZone{{Domain: "netbird.cloud", Records: []nbdns.SimpleRecord{{Name: "netbird.cloud", Type: 1, Class: nbdns.DefaultClass, TTL: 300, RData: "10.0.0.1"}}}},
+			initUpstreamMap: []handlerWrapper{
+				{
+					domain:   "netbird.cloud",
+					handler:  &mockHandler{},
+					priority: PriorityUpstream,
+				},
+			},
+			initSerial:  0,
+			inputSerial: 1,
+			inputUpdate: nbdns.Config{
+				ServiceEnable: true,
+				CustomZones: []nbdns.CustomZone{
+					{
+						Domain:  "netbird.cloud",
+						Records: zoneRecords,
+					},
+				},
+				NameServerGroups: []*nbdns.NameServerGroup{
+					{
+						Domains:     []string{"netbird.io"},
+						NameServers: nameServers,
+					},
+				},
+			},
+			expectedUpstreamMap: []handlerWrapper{
+				{
+					domain:   "netbird.io",
+					priority: PriorityUpstream,
+				},
+				{
+					domain:   "netbird.cloud",
+					priority: PriorityLocal,
+				},
+			},
+			expectedLocalQs: []dns.Question{{Name: zoneRecords[0].Name, Qtype: 1, Qclass: 1}},
+		},
+		{
+			name:            "Smaller Config Serial Should Be Skipped",
+			initLocalZones:  []nbdns.CustomZone{},
+			initUpstreamMap: nil,
+			initSerial:      2,
+			inputSerial:     1,
+			shouldFail:      true,
+		},
+		{
+			name:            "Empty NS Group Domain Or Not Primary Element Should Fail",
+			initLocalZones:  []nbdns.CustomZone{},
+			initUpstreamMap: nil,
+			initSerial:      0,
+			inputSerial:     1,
+			inputUpdate: nbdns.Config{
+				ServiceEnable: true,
+				CustomZones: []nbdns.CustomZone{
+					{
+						Domain:  "netbird.cloud",
+						Records: zoneRecords,
+					},
+				},
+				NameServerGroups: []*nbdns.NameServerGroup{
+					{
+						NameServers: nameServers,
+					},
+				},
+			},
+			shouldFail: true,
+		},
+		{
+			name:            "Invalid NS Group Nameservers list Should Fail",
+			initLocalZones:  []nbdns.CustomZone{},
+			initUpstreamMap: nil,
+			initSerial:      0,
+			inputSerial:     1,
+			inputUpdate: nbdns.Config{
+				ServiceEnable: true,
+				CustomZones: []nbdns.CustomZone{
+					{
+						Domain:  "netbird.cloud",
+						Records: zoneRecords,
+					},
+				},
+				NameServerGroups: []*nbdns.NameServerGroup{
+					{
+						NameServers: nameServers,
+					},
+				},
+			},
+			shouldFail: true,
+		},
+		{
+			name:            "Invalid Custom Zone Records list Should Skip",
+			initLocalZones:  []nbdns.CustomZone{},
+			initUpstreamMap: nil,
+			initSerial:      0,
+			inputSerial:     1,
+			inputUpdate: nbdns.Config{
+				ServiceEnable: true,
+				CustomZones: []nbdns.CustomZone{
+					{
+						Domain: "netbird.cloud",
+					},
+				},
+				NameServerGroups: []*nbdns.NameServerGroup{
+					{
+						NameServers: nameServers,
+						Primary:     true,
+					},
+				},
+			},
+			expectedUpstreamMap: []handlerWrapper{{
+				domain:   ".",
+				priority: PriorityDefault,
+			}},
+		},
+		{
+			name:           "Empty Config Should Succeed and Clean Maps",
+			initLocalZones: []nbdns.CustomZone{{Domain: "netbird.cloud", Records: []nbdns.SimpleRecord{{Name: "netbird.cloud", Type: int(dns.TypeA), Class: nbdns.DefaultClass, TTL: 300, RData: "10.0.0.1"}}}},
+			initUpstreamMap: []handlerWrapper{
+				{
+					domain:   zoneRecords[0].Name,
+					handler:  &mockHandler{},
+					priority: PriorityUpstream,
+				},
+			},
+			initSerial:          0,
+			inputSerial:         1,
+			inputUpdate:         nbdns.Config{ServiceEnable: true},
+			expectedUpstreamMap: nil,
+			expectedLocalQs:     []dns.Question{},
+		},
+		{
+			name:           "Disabled Service Should clean map",
+			initLocalZones: []nbdns.CustomZone{{Domain: "netbird.cloud", Records: []nbdns.SimpleRecord{{Name: "netbird.cloud", Type: int(dns.TypeA), Class: nbdns.DefaultClass, TTL: 300, RData: "10.0.0.1"}}}},
+			initUpstreamMap: []handlerWrapper{
+				{
+					domain:   zoneRecords[0].Name,
+					handler:  &mockHandler{},
+					priority: PriorityUpstream,
+				},
+			},
+			initSerial:          0,
+			inputSerial:         1,
+			inputUpdate:         nbdns.Config{ServiceEnable: false},
+			expectedUpstreamMap: nil,
+			expectedLocalQs:     []dns.Question{},
+		},
+	}
+
+	for n, testCase := range testCases {
+		t.Run(testCase.name, func(t *testing.T) {
+			privKey, _ := wgtypes.GenerateKey()
+			newNet, err := stdnet.NewNet(context.Background(), nil)
+			if err != nil {
+				t.Fatal(err)
+			}
+
+			opts := iface.WGIFaceOpts{
+				IFaceName:    fmt.Sprintf("utun230%d", n),
+				Address:      wgaddr.MustParseWGAddress(fmt.Sprintf("100.66.100.%d/32", n+1)),
+				WGPort:       33100,
+				WGPrivKey:    privKey.String(),
+				MTU:          iface.DefaultMTU,
+				TransportNet: newNet,
+			}
+
+			wgIface, err := iface.NewWGIFace(opts)
+			if err != nil {
+				t.Fatal(err)
+			}
+			err = wgIface.Create()
+			if err != nil {
+				t.Fatal(err)
+			}
+			defer func() {
+				err = wgIface.Close()
+				if err != nil {
+					t.Log(err)
+				}
+			}()
+			dnsServer, err := NewDefaultServer(context.Background(), DefaultServerConfig{
+				WgInterface:    wgIface,
+				CustomAddress:  "",
+				StatusRecorder: peer.NewRecorder("mgm"),
+				StateManager:   nil,
+				DisableSys:     false,
+			})
+			if err != nil {
+				t.Fatal(err)
+			}
+			err = dnsServer.Initialize()
+			if err != nil {
+				t.Fatal(err)
+			}
+			defer func() {
+				err = dnsServer.hostManager.restoreHostDNS()
+				if err != nil {
+					t.Log(err)
+				}
+			}()
+
+			dnsServer.dnsMuxHandlers = testCase.initUpstreamMap
+			dnsServer.localResolver.Update(testCase.initLocalZones)
+			dnsServer.updateSerial = testCase.initSerial
+
+			err = dnsServer.UpdateDNSServer(testCase.inputSerial, testCase.inputUpdate)
+			if err != nil {
+				if testCase.shouldFail {
+					return
+				}
+				t.Fatalf("update dns server should not fail, got error: %v", err)
+			}
+
+			if len(dnsServer.dnsMuxHandlers) != len(testCase.expectedUpstreamMap) {
+				t.Fatalf("update upstream failed, map size is different than expected, want %d, got %d", len(testCase.expectedUpstreamMap), len(dnsServer.dnsMuxHandlers))
+			}
+
+			for _, expected := range testCase.expectedUpstreamMap {
+				found := false
+				for _, got := range dnsServer.dnsMuxHandlers {
+					if got.domain == expected.domain && got.priority == expected.priority {
+						found = true
+						break
+					}
+				}
+				if !found {
+					t.Fatalf("update upstream failed, handler for domain=%s priority=%d not found in dnsMuxHandlers: %#v", expected.domain, expected.priority, dnsServer.dnsMuxHandlers)
+				}
+			}
+
+			var responseMSG *dns.Msg
+			responseWriter := &test.MockResponseWriter{
+				WriteMsgFunc: func(m *dns.Msg) error {
+					responseMSG = m
+					return nil
+				},
+			}
+			for _, q := range testCase.expectedLocalQs {
+				dnsServer.localResolver.ServeDNS(responseWriter, &dns.Msg{
+					Question: []dns.Question{q},
+				})
+			}
+
+			if len(testCase.expectedLocalQs) > 0 {
+				assert.NotNil(t, responseMSG, "response message should not be nil")
+				assert.Equal(t, dns.RcodeSuccess, responseMSG.Rcode, "response code should be success")
+				assert.NotEmpty(t, responseMSG.Answer, "response message should have answers")
+			}
+		})
+	}
+}
+
+func TestDNSFakeResolverHandleUpdates(t *testing.T) {
+	ov := os.Getenv("NB_WG_KERNEL_DISABLED")
+	defer t.Setenv("NB_WG_KERNEL_DISABLED", ov)
+
+	t.Setenv("NB_WG_KERNEL_DISABLED", "true")
+	newNet, err := stdnet.NewNet(context.Background(), []string{"utun2301"})
+	if err != nil {
+		t.Errorf("create stdnet: %v", err)
+		return
+	}
+
+	privKey, _ := wgtypes.GeneratePrivateKey()
+	opts := iface.WGIFaceOpts{
+		IFaceName:    "utun2301",
+		Address:      wgaddr.MustParseWGAddress("100.66.100.1/32"),
+		WGPort:       33100,
+		WGPrivKey:    privKey.String(),
+		MTU:          iface.DefaultMTU,
+		TransportNet: newNet,
+	}
+	wgIface, err := iface.NewWGIFace(opts)
+	if err != nil {
+		t.Errorf("build interface wireguard: %v", err)
+		return
+	}
+
+	err = wgIface.Create()
+	if err != nil {
+		t.Errorf("create and init wireguard interface: %v", err)
+		return
+	}
+	defer func() {
+		if err = wgIface.Close(); err != nil {
+			t.Logf("close wireguard interface: %v", err)
+		}
+	}()
+
+	ctrl := gomock.NewController(t)
+	defer ctrl.Finish()
+
+	packetfilter := pfmock.NewMockPacketFilter(ctrl)
+	packetfilter.EXPECT().FilterOutbound(gomock.Any(), gomock.Any()).AnyTimes()
+	packetfilter.EXPECT().SetUDPPacketHook(gomock.Any(), gomock.Any(), gomock.Any()).AnyTimes()
+	packetfilter.EXPECT().SetTCPPacketHook(gomock.Any(), gomock.Any(), gomock.Any()).AnyTimes()
+
+	if err := wgIface.SetFilter(packetfilter); err != nil {
+		t.Errorf("set packet filter: %v", err)
+		return
+	}
+
+	dnsServer, err := NewDefaultServer(context.Background(), DefaultServerConfig{
+		WgInterface:    wgIface,
+		CustomAddress:  "",
+		StatusRecorder: peer.NewRecorder("mgm"),
+		StateManager:   nil,
+		DisableSys:     false,
+	})
+	if err != nil {
+		t.Errorf("create DNS server: %v", err)
+		return
+	}
+
+	err = dnsServer.Initialize()
+	if err != nil {
+		t.Errorf("run DNS server: %v", err)
+		return
+	}
+	defer func() {
+		if err = dnsServer.hostManager.restoreHostDNS(); err != nil {
+			t.Logf("restore DNS settings on the host: %v", err)
+			return
+		}
+	}()
+
+	dnsServer.dnsMuxHandlers = []handlerWrapper{
+		{
+			domain:   zoneRecords[0].Name,
+			handler:  &local.Resolver{},
+			priority: PriorityUpstream,
+		},
+	}
+	dnsServer.localResolver.Update([]nbdns.CustomZone{{Domain: "netbird.cloud", Records: []nbdns.SimpleRecord{{Name: "netbird.cloud", Type: int(dns.TypeA), Class: nbdns.DefaultClass, TTL: 300, RData: "10.0.0.1"}}}})
+	dnsServer.updateSerial = 0
+
+	nameServers := []nbdns.NameServer{
+		{
+			IP:     netip.MustParseAddr("8.8.8.8"),
+			NSType: nbdns.UDPNameServerType,
+			Port:   53,
+		},
+		{
+			IP:     netip.MustParseAddr("8.8.4.4"),
+			NSType: nbdns.UDPNameServerType,
+			Port:   53,
+		},
+	}
+
+	update := nbdns.Config{
+		ServiceEnable: true,
+		CustomZones: []nbdns.CustomZone{
+			{
+				Domain:  "netbird.cloud",
+				Records: zoneRecords,
+			},
+		},
+		NameServerGroups: []*nbdns.NameServerGroup{
+			{
+				Domains:     []string{"netbird.io"},
+				NameServers: nameServers,
+			},
+			{
+				NameServers: nameServers,
+				Primary:     true,
+			},
+		},
+	}
+
+	// Start the server with regular configuration
+	if err := dnsServer.UpdateDNSServer(1, update); err != nil {
+		t.Fatalf("update dns server should not fail, got error: %v", err)
+		return
+	}
+
+	update2 := update
+	update2.ServiceEnable = false
+	// Disable the server, stop the listener
+	if err := dnsServer.UpdateDNSServer(2, update2); err != nil {
+		t.Fatalf("update dns server should not fail, got error: %v", err)
+		return
+	}
+
+	update3 := update2
+	update3.NameServerGroups = update3.NameServerGroups[:1]
+	// But service still get updates and we checking that we handle
+	// internal state in the right way
+	if err := dnsServer.UpdateDNSServer(3, update3); err != nil {
+		t.Fatalf("update dns server should not fail, got error: %v", err)
+		return
+	}
+}

client/internal/dns/server_test.go

@@ -10,7 +10,6 @@ import (
 	"testing"
 	"time"
 
-	"github.com/golang/mock/gomock"
 	"github.com/miekg/dns"
 	log "github.com/sirupsen/logrus"
 	"github.com/stretchr/testify/assert"
@@ -23,7 +22,6 @@ import (
 	"github.com/netbirdio/netbird/client/iface"
 	"github.com/netbirdio/netbird/client/iface/configurer"
 	"github.com/netbirdio/netbird/client/iface/device"
-	pfmock "github.com/netbirdio/netbird/client/iface/mocks"
 	"github.com/netbirdio/netbird/client/iface/wgaddr"
 	"github.com/netbirdio/netbird/client/internal/dns/local"
 	"github.com/netbirdio/netbird/client/internal/dns/test"
@@ -104,466 +102,6 @@ func init() {
 	formatter.SetTextFormatter(log.StandardLogger())
 }
 
-func TestUpdateDNSServer(t *testing.T) {
-
-	nameServers := []nbdns.NameServer{
-		{
-			IP:     netip.MustParseAddr("8.8.8.8"),
-			NSType: nbdns.UDPNameServerType,
-			Port:   53,
-		},
-		{
-			IP:     netip.MustParseAddr("8.8.4.4"),
-			NSType: nbdns.UDPNameServerType,
-			Port:   53,
-		},
-	}
-
-	testCases := []struct {
-		name                string
-		initUpstreamMap     []handlerWrapper
-		initLocalZones      []nbdns.CustomZone
-		initSerial          uint64
-		inputSerial         uint64
-		inputUpdate         nbdns.Config
-		shouldFail          bool
-		expectedUpstreamMap []handlerWrapper
-		expectedLocalQs     []dns.Question
-	}{
-		{
-			name:            "Initial Config Should Succeed",
-			initUpstreamMap: nil,
-			initSerial:      0,
-			inputSerial:     1,
-			inputUpdate: nbdns.Config{
-				ServiceEnable: true,
-				CustomZones: []nbdns.CustomZone{
-					{
-						Domain:  "netbird.cloud",
-						Records: zoneRecords,
-					},
-				},
-				NameServerGroups: []*nbdns.NameServerGroup{
-					{
-						Domains:     []string{"netbird.io"},
-						NameServers: nameServers,
-					},
-					{
-						NameServers: nameServers,
-						Primary:     true,
-					},
-				},
-			},
-			expectedUpstreamMap: []handlerWrapper{
-				{
-					domain:   "netbird.io",
-					priority: PriorityUpstream,
-				},
-				{
-					domain:   "netbird.cloud",
-					priority: PriorityLocal,
-				},
-				{
-					domain:   nbdns.RootZone,
-					priority: PriorityDefault,
-				},
-			},
-			expectedLocalQs: []dns.Question{{Name: "peera.netbird.cloud.", Qtype: dns.TypeA, Qclass: dns.ClassINET}},
-		},
-		{
-			name:           "New Config Should Succeed",
-			initLocalZones: []nbdns.CustomZone{{Domain: "netbird.cloud", Records: []nbdns.SimpleRecord{{Name: "netbird.cloud", Type: 1, Class: nbdns.DefaultClass, TTL: 300, RData: "10.0.0.1"}}}},
-			initUpstreamMap: []handlerWrapper{
-				{
-					domain:   "netbird.cloud",
-					handler:  &mockHandler{},
-					priority: PriorityUpstream,
-				},
-			},
-			initSerial:  0,
-			inputSerial: 1,
-			inputUpdate: nbdns.Config{
-				ServiceEnable: true,
-				CustomZones: []nbdns.CustomZone{
-					{
-						Domain:  "netbird.cloud",
-						Records: zoneRecords,
-					},
-				},
-				NameServerGroups: []*nbdns.NameServerGroup{
-					{
-						Domains:     []string{"netbird.io"},
-						NameServers: nameServers,
-					},
-				},
-			},
-			expectedUpstreamMap: []handlerWrapper{
-				{
-					domain:   "netbird.io",
-					priority: PriorityUpstream,
-				},
-				{
-					domain:   "netbird.cloud",
-					priority: PriorityLocal,
-				},
-			},
-			expectedLocalQs: []dns.Question{{Name: zoneRecords[0].Name, Qtype: 1, Qclass: 1}},
-		},
-		{
-			name:            "Smaller Config Serial Should Be Skipped",
-			initLocalZones:  []nbdns.CustomZone{},
-			initUpstreamMap: nil,
-			initSerial:      2,
-			inputSerial:     1,
-			shouldFail:      true,
-		},
-		{
-			name:            "Empty NS Group Domain Or Not Primary Element Should Fail",
-			initLocalZones:  []nbdns.CustomZone{},
-			initUpstreamMap: nil,
-			initSerial:      0,
-			inputSerial:     1,
-			inputUpdate: nbdns.Config{
-				ServiceEnable: true,
-				CustomZones: []nbdns.CustomZone{
-					{
-						Domain:  "netbird.cloud",
-						Records: zoneRecords,
-					},
-				},
-				NameServerGroups: []*nbdns.NameServerGroup{
-					{
-						NameServers: nameServers,
-					},
-				},
-			},
-			shouldFail: true,
-		},
-		{
-			name:            "Invalid NS Group Nameservers list Should Fail",
-			initLocalZones:  []nbdns.CustomZone{},
-			initUpstreamMap: nil,
-			initSerial:      0,
-			inputSerial:     1,
-			inputUpdate: nbdns.Config{
-				ServiceEnable: true,
-				CustomZones: []nbdns.CustomZone{
-					{
-						Domain:  "netbird.cloud",
-						Records: zoneRecords,
-					},
-				},
-				NameServerGroups: []*nbdns.NameServerGroup{
-					{
-						NameServers: nameServers,
-					},
-				},
-			},
-			shouldFail: true,
-		},
-		{
-			name:            "Invalid Custom Zone Records list Should Skip",
-			initLocalZones:  []nbdns.CustomZone{},
-			initUpstreamMap: nil,
-			initSerial:      0,
-			inputSerial:     1,
-			inputUpdate: nbdns.Config{
-				ServiceEnable: true,
-				CustomZones: []nbdns.CustomZone{
-					{
-						Domain: "netbird.cloud",
-					},
-				},
-				NameServerGroups: []*nbdns.NameServerGroup{
-					{
-						NameServers: nameServers,
-						Primary:     true,
-					},
-				},
-			},
-			expectedUpstreamMap: []handlerWrapper{{
-				domain:   ".",
-				priority: PriorityDefault,
-			}},
-		},
-		{
-			name:           "Empty Config Should Succeed and Clean Maps",
-			initLocalZones: []nbdns.CustomZone{{Domain: "netbird.cloud", Records: []nbdns.SimpleRecord{{Name: "netbird.cloud", Type: int(dns.TypeA), Class: nbdns.DefaultClass, TTL: 300, RData: "10.0.0.1"}}}},
-			initUpstreamMap: []handlerWrapper{
-				{
-					domain:   zoneRecords[0].Name,
-					handler:  &mockHandler{},
-					priority: PriorityUpstream,
-				},
-			},
-			initSerial:          0,
-			inputSerial:         1,
-			inputUpdate:         nbdns.Config{ServiceEnable: true},
-			expectedUpstreamMap: nil,
-			expectedLocalQs:     []dns.Question{},
-		},
-		{
-			name:           "Disabled Service Should clean map",
-			initLocalZones: []nbdns.CustomZone{{Domain: "netbird.cloud", Records: []nbdns.SimpleRecord{{Name: "netbird.cloud", Type: int(dns.TypeA), Class: nbdns.DefaultClass, TTL: 300, RData: "10.0.0.1"}}}},
-			initUpstreamMap: []handlerWrapper{
-				{
-					domain:   zoneRecords[0].Name,
-					handler:  &mockHandler{},
-					priority: PriorityUpstream,
-				},
-			},
-			initSerial:          0,
-			inputSerial:         1,
-			inputUpdate:         nbdns.Config{ServiceEnable: false},
-			expectedUpstreamMap: nil,
-			expectedLocalQs:     []dns.Question{},
-		},
-	}
-
-	for n, testCase := range testCases {
-		t.Run(testCase.name, func(t *testing.T) {
-			privKey, _ := wgtypes.GenerateKey()
-			newNet, err := stdnet.NewNet(context.Background(), nil)
-			if err != nil {
-				t.Fatal(err)
-			}
-
-			opts := iface.WGIFaceOpts{
-				IFaceName:    fmt.Sprintf("utun230%d", n),
-				Address:      wgaddr.MustParseWGAddress(fmt.Sprintf("100.66.100.%d/32", n+1)),
-				WGPort:       33100,
-				WGPrivKey:    privKey.String(),
-				MTU:          iface.DefaultMTU,
-				TransportNet: newNet,
-			}
-
-			wgIface, err := iface.NewWGIFace(opts)
-			if err != nil {
-				t.Fatal(err)
-			}
-			err = wgIface.Create()
-			if err != nil {
-				t.Fatal(err)
-			}
-			defer func() {
-				err = wgIface.Close()
-				if err != nil {
-					t.Log(err)
-				}
-			}()
-			dnsServer, err := NewDefaultServer(context.Background(), DefaultServerConfig{
-				WgInterface:    wgIface,
-				CustomAddress:  "",
-				StatusRecorder: peer.NewRecorder("mgm"),
-				StateManager:   nil,
-				DisableSys:     false,
-			})
-			if err != nil {
-				t.Fatal(err)
-			}
-			err = dnsServer.Initialize()
-			if err != nil {
-				t.Fatal(err)
-			}
-			defer func() {
-				err = dnsServer.hostManager.restoreHostDNS()
-				if err != nil {
-					t.Log(err)
-				}
-			}()
-
-			dnsServer.dnsMuxHandlers = testCase.initUpstreamMap
-			dnsServer.localResolver.Update(testCase.initLocalZones)
-			dnsServer.updateSerial = testCase.initSerial
-
-			err = dnsServer.UpdateDNSServer(testCase.inputSerial, testCase.inputUpdate)
-			if err != nil {
-				if testCase.shouldFail {
-					return
-				}
-				t.Fatalf("update dns server should not fail, got error: %v", err)
-			}
-
-			if len(dnsServer.dnsMuxHandlers) != len(testCase.expectedUpstreamMap) {
-				t.Fatalf("update upstream failed, map size is different than expected, want %d, got %d", len(testCase.expectedUpstreamMap), len(dnsServer.dnsMuxHandlers))
-			}
-
-			for _, expected := range testCase.expectedUpstreamMap {
-				found := false
-				for _, got := range dnsServer.dnsMuxHandlers {
-					if got.domain == expected.domain && got.priority == expected.priority {
-						found = true
-						break
-					}
-				}
-				if !found {
-					t.Fatalf("update upstream failed, handler for domain=%s priority=%d not found in dnsMuxHandlers: %#v", expected.domain, expected.priority, dnsServer.dnsMuxHandlers)
-				}
-			}
-
-			var responseMSG *dns.Msg
-			responseWriter := &test.MockResponseWriter{
-				WriteMsgFunc: func(m *dns.Msg) error {
-					responseMSG = m
-					return nil
-				},
-			}
-			for _, q := range testCase.expectedLocalQs {
-				dnsServer.localResolver.ServeDNS(responseWriter, &dns.Msg{
-					Question: []dns.Question{q},
-				})
-			}
-
-			if len(testCase.expectedLocalQs) > 0 {
-				assert.NotNil(t, responseMSG, "response message should not be nil")
-				assert.Equal(t, dns.RcodeSuccess, responseMSG.Rcode, "response code should be success")
-				assert.NotEmpty(t, responseMSG.Answer, "response message should have answers")
-			}
-		})
-	}
-}
-
-func TestDNSFakeResolverHandleUpdates(t *testing.T) {
-	ov := os.Getenv("NB_WG_KERNEL_DISABLED")
-	defer t.Setenv("NB_WG_KERNEL_DISABLED", ov)
-
-	t.Setenv("NB_WG_KERNEL_DISABLED", "true")
-	newNet, err := stdnet.NewNet(context.Background(), []string{"utun2301"})
-	if err != nil {
-		t.Errorf("create stdnet: %v", err)
-		return
-	}
-
-	privKey, _ := wgtypes.GeneratePrivateKey()
-	opts := iface.WGIFaceOpts{
-		IFaceName:    "utun2301",
-		Address:      wgaddr.MustParseWGAddress("100.66.100.1/32"),
-		WGPort:       33100,
-		WGPrivKey:    privKey.String(),
-		MTU:          iface.DefaultMTU,
-		TransportNet: newNet,
-	}
-	wgIface, err := iface.NewWGIFace(opts)
-	if err != nil {
-		t.Errorf("build interface wireguard: %v", err)
-		return
-	}
-
-	err = wgIface.Create()
-	if err != nil {
-		t.Errorf("create and init wireguard interface: %v", err)
-		return
-	}
-	defer func() {
-		if err = wgIface.Close(); err != nil {
-			t.Logf("close wireguard interface: %v", err)
-		}
-	}()
-
-	ctrl := gomock.NewController(t)
-	defer ctrl.Finish()
-
-	packetfilter := pfmock.NewMockPacketFilter(ctrl)
-	packetfilter.EXPECT().FilterOutbound(gomock.Any(), gomock.Any()).AnyTimes()
-	packetfilter.EXPECT().SetUDPPacketHook(gomock.Any(), gomock.Any(), gomock.Any()).AnyTimes()
-	packetfilter.EXPECT().SetTCPPacketHook(gomock.Any(), gomock.Any(), gomock.Any()).AnyTimes()
-
-	if err := wgIface.SetFilter(packetfilter); err != nil {
-		t.Errorf("set packet filter: %v", err)
-		return
-	}
-
-	dnsServer, err := NewDefaultServer(context.Background(), DefaultServerConfig{
-		WgInterface:    wgIface,
-		CustomAddress:  "",
-		StatusRecorder: peer.NewRecorder("mgm"),
-		StateManager:   nil,
-		DisableSys:     false,
-	})
-	if err != nil {
-		t.Errorf("create DNS server: %v", err)
-		return
-	}
-
-	err = dnsServer.Initialize()
-	if err != nil {
-		t.Errorf("run DNS server: %v", err)
-		return
-	}
-	defer func() {
-		if err = dnsServer.hostManager.restoreHostDNS(); err != nil {
-			t.Logf("restore DNS settings on the host: %v", err)
-			return
-		}
-	}()
-
-	dnsServer.dnsMuxHandlers = []handlerWrapper{
-		{
-			domain:   zoneRecords[0].Name,
-			handler:  &local.Resolver{},
-			priority: PriorityUpstream,
-		},
-	}
-	dnsServer.localResolver.Update([]nbdns.CustomZone{{Domain: "netbird.cloud", Records: []nbdns.SimpleRecord{{Name: "netbird.cloud", Type: int(dns.TypeA), Class: nbdns.DefaultClass, TTL: 300, RData: "10.0.0.1"}}}})
-	dnsServer.updateSerial = 0
-
-	nameServers := []nbdns.NameServer{
-		{
-			IP:     netip.MustParseAddr("8.8.8.8"),
-			NSType: nbdns.UDPNameServerType,
-			Port:   53,
-		},
-		{
-			IP:     netip.MustParseAddr("8.8.4.4"),
-			NSType: nbdns.UDPNameServerType,
-			Port:   53,
-		},
-	}
-
-	update := nbdns.Config{
-		ServiceEnable: true,
-		CustomZones: []nbdns.CustomZone{
-			{
-				Domain:  "netbird.cloud",
-				Records: zoneRecords,
-			},
-		},
-		NameServerGroups: []*nbdns.NameServerGroup{
-			{
-				Domains:     []string{"netbird.io"},
-				NameServers: nameServers,
-			},
-			{
-				NameServers: nameServers,
-				Primary:     true,
-			},
-		},
-	}
-
-	// Start the server with regular configuration
-	if err := dnsServer.UpdateDNSServer(1, update); err != nil {
-		t.Fatalf("update dns server should not fail, got error: %v", err)
-		return
-	}
-
-	update2 := update
-	update2.ServiceEnable = false
-	// Disable the server, stop the listener
-	if err := dnsServer.UpdateDNSServer(2, update2); err != nil {
-		t.Fatalf("update dns server should not fail, got error: %v", err)
-		return
-	}
-
-	update3 := update2
-	update3.NameServerGroups = update3.NameServerGroups[:1]
-	// But service still get updates and we checking that we handle
-	// internal state in the right way
-	if err := dnsServer.UpdateDNSServer(3, update3); err != nil {
-		t.Fatalf("update dns server should not fail, got error: %v", err)
-		return
-	}
-}
-
 func TestDNSServerStartStop(t *testing.T) {
 	testCases := []struct {
 		name     string

client/internal/engine.go

@@ -210,6 +210,12 @@ type Engine struct {
 	// networkSerial is the latest CurrentSerial (state ID) of the network sent by the Management service
 	networkSerial uint64
 
+	// forwardingRules holds the ingress forward rules applied for the current target.
+	// Wholesale sections (incl. forward rules) run only on the first pass of a target;
+	// it is stashed here so the final, peer-converged pass can build the lazy-connection
+	// exclude list without recomputing them on every bounded peer pass.
+	forwardingRules []firewallManager.ForwardRule
+
 	networkMonitor *networkmonitor.NetworkMonitor
 
 	sshServer sshServer
@@ -762,7 +768,15 @@ func (e *Engine) blockLanAccess() {
 
 // modifyPeers updates peers that have been modified (e.g. IP address has been changed).
 // It closes the existing connection, removes it from the peerConns map, and creates a new one.
-func (e *Engine) modifyPeers(peersUpdate []*mgmProto.RemotePeerConfig) error {
+// maxPeersPerSyncPass is the default per-pass cap on how many peers each of
+// removePeers/modifyPeers/addNewPeers applies, so syncMsgMux is held only for a
+// batch at a time and other subsystems can interleave between passes. It is
+// passed in (not read globally) so tests can exercise the multi-pass path.
+const maxPeersPerSyncPass = 300
+
+// modifyPeers re-applies up to maxBatch changed peers per call. It returns true
+// when more changed peers remained than the cap, so the caller re-runs.
+func (e *Engine) modifyPeers(peersUpdate []*mgmProto.RemotePeerConfig, maxBatch int) (bool, error) {
 
 	// first, check if peers have been modified
 	var modified []*mgmProto.RemotePeerConfig
@@ -792,26 +806,32 @@ func (e *Engine) modifyPeers(peersUpdate []*mgmProto.RemotePeerConfig) error {
 		}
 	}
 
+	more := false
+	if len(modified) > maxBatch {
+		modified = modified[:maxBatch]
+		more = true
+	}
+
 	// second, close all modified connections and remove them from the state map
 	for _, p := range modified {
-		err := e.removePeer(p.GetWgPubKey())
+		if err := e.removePeer(p.GetWgPubKey()); err != nil {
-		if err != nil {
+			return false, err
-			return err
 		}
 	}
 	// third, add the peer connections again
 	for _, p := range modified {
-		err := e.addNewPeer(p)
+		if err := e.addNewPeer(p); err != nil {
-		if err != nil {
+			return false, err
-			return err
 		}
 	}
-	return nil
+	return more, nil
 }
 
 // removePeers finds and removes peers that do not exist anymore in the network map received from the Management Service.
 // It also removes peers that have been modified (e.g. change of IP address). They will be added again in addPeers method.
-func (e *Engine) removePeers(peersUpdate []*mgmProto.RemotePeerConfig) error {
+// removePeers removes up to maxBatch peers per call. It returns true when more
+// peers remained to remove than the cap, so the caller re-runs.
+func (e *Engine) removePeers(peersUpdate []*mgmProto.RemotePeerConfig, maxBatch int) (bool, error) {
 	newPeers := make([]string, 0, len(peersUpdate))
 	for _, p := range peersUpdate {
 		newPeers = append(newPeers, p.GetWgPubKey())
@@ -819,14 +839,19 @@ func (e *Engine) removePeers(peersUpdate []*mgmProto.RemotePeerConfig) error {
 
 	toRemove := util.SliceDiff(e.peerStore.PeersPubKey(), newPeers)
 
+	more := false
+	if len(toRemove) > maxBatch {
+		toRemove = toRemove[:maxBatch]
+		more = true
+	}
+
 	for _, p := range toRemove {
-		err := e.removePeer(p)
+		if err := e.removePeer(p); err != nil {
-		if err != nil {
+			return false, err
-			return err
 		}
 		log.Infof("removed peer %s", p)
 	}
-	return nil
+	return more, nil
 }
 
 func (e *Engine) removeAllPeers() error {
@@ -895,19 +920,17 @@ func (e *Engine) handleAutoUpdateVersion(autoUpdateSettings *mgmProto.AutoUpdate
 	e.updateManager.SetVersion(autoUpdateSettings.Version, autoUpdateSettings.AlwaysUpdate)
 }
 
-func (e *Engine) handleSync(update *mgmProto.SyncResponse) error {
+// applySyncPass applies one bounded pass of the sync update under syncMsgMux and
-	started := time.Now()
+// returns true if more peers remained than the per-pass cap. It is driven by the
-	defer func() {
+// mapStateManager, which re-invokes it (releasing the lock between passes) until
-		duration := time.Since(started)
+// the update is fully applied.
-		log.Infof("sync finished in %s", duration)
+func (e *Engine) applySyncPass(update *mgmProto.SyncResponse, firstPass bool) (bool, error) {
-		e.clientMetrics.RecordSyncDuration(e.ctx, duration)
-	}()
 	e.syncMsgMux.Lock()
 	defer e.syncMsgMux.Unlock()
 
 	// Check context INSIDE lock to ensure atomicity with shutdown
 	if e.ctx.Err() != nil {
-		return e.ctx.Err()
+		return false, e.ctx.Err()
 	}
 
 	if update.NetworkMap != nil && update.NetworkMap.PeerConfig != nil {
@@ -915,7 +938,7 @@ func (e *Engine) handleSync(update *mgmProto.SyncResponse) error {
 	}
 
 	if err := e.updateNetbirdConfig(update.GetNetbirdConfig()); err != nil {
-		return err
+		return false, err
 	}
 
 	// Posture checks are bound to the network map presence:
@@ -925,23 +948,22 @@ func (e *Engine) handleSync(update *mgmProto.SyncResponse) error {
 	//                                        leave the previously applied checks untouched
 	nm := update.GetNetworkMap()
 	if nm == nil {
-		return nil
+		return false, nil
 	}
 
 	if err := e.updateChecksIfNew(update.Checks); err != nil {
-		return err
+		return false, err
 	}
 
-	e.persistSyncResponse(update)
-
 	// only apply new changes and ignore old ones
-	if err := e.updateNetworkMap(nm); err != nil {
+	more, err := e.updateNetworkMap(nm, maxPeersPerSyncPass, firstPass)
-		return err
+	if err != nil {
+		return false, err
 	}
 
 	e.statusRecorder.PublishEvent(cProto.SystemEvent_INFO, cProto.SystemEvent_SYSTEM, "Network map updated", "", nil)
 
-	return nil
+	return more, nil
 }
 
 // updateNetbirdConfig applies the management-provided NetBird configuration:
@@ -987,6 +1009,13 @@ func (e *Engine) updateNetbirdConfig(wCfg *mgmProto.NetbirdConfig) error {
 // (not syncMsgMux) is held for the whole Set so the store cannot be cleared (disabled /
 // engine close) mid-call and have this write resurrect a file that was just removed.
 func (e *Engine) persistSyncResponse(update *mgmProto.SyncResponse) {
+	// Only persist updates that carry a network map. Config-only updates (e.g. relay
+	// token rotation, STUN/TURN) have a nil NetworkMap; persisting them would overwrite
+	// the last full map on disk and break restore-on-restart.
+	if update.GetNetworkMap() == nil {
+		return
+	}
+
 	e.syncRespMux.RLock()
 	defer e.syncRespMux.RUnlock()
 
@@ -1066,7 +1095,7 @@ func (e *Engine) updateChecksIfNew(checks []*mgmProto.Checks) error {
 	}
 	e.checks = checks
 
-	info, err := system.GetInfoWithChecks(e.ctx, checks)
+	info, err := system.GetInfoWithChecks(e.ctx, checks, e.overlayAddresses()...)
 	if err != nil {
 		log.Warnf("failed to get system info with checks: %v", err)
 		info = system.GetInfo(e.ctx)
@@ -1097,6 +1126,20 @@ func (e *Engine) updateChecksIfNew(checks []*mgmProto.Checks) error {
 	return nil
 }
 
+// overlayAddresses returns our own WireGuard overlay address (v4 and v6) so it
+// can be excluded from the reported network addresses; the interface coming and
+// going otherwise churns the peer meta on the management server.
+func (e *Engine) overlayAddresses() []netip.Addr {
+	var ips []netip.Addr
+	if e.config.WgAddr.IP.IsValid() {
+		ips = append(ips, e.config.WgAddr.IP)
+	}
+	if e.config.WgAddr.HasIPv6() {
+		ips = append(ips, e.config.WgAddr.IPv6)
+	}
+	return ips
+}
+
 func (e *Engine) updateConfig(conf *mgmProto.PeerConfig) error {
 	if e.wgInterface == nil {
 		return errors.New("wireguard interface is not initialized")
@@ -1240,7 +1283,7 @@ func (e *Engine) receiveManagementEvents() {
 	e.shutdownWg.Add(1)
 	go func() {
 		defer e.shutdownWg.Done()
-		info, err := system.GetInfoWithChecks(e.ctx, e.checks)
+		info, err := system.GetInfoWithChecks(e.ctx, e.checks, e.overlayAddresses()...)
 		if err != nil {
 			log.Warnf("failed to get system info with checks: %v", err)
 			info = system.GetInfo(e.ctx)
@@ -1264,7 +1307,19 @@ func (e *Engine) receiveManagementEvents() {
 			e.config.DisableSSHAuth,
 		)
 
-		err = e.mgmClient.Sync(e.ctx, info, e.handleSync)
+		// The map-state manager converges the latest update in the background in
+		// bounded passes; the stream callback only hands it the newest target.
+		manager := newMapStateManager(e.applySyncPass, e.persistSyncResponse, func(d time.Duration) {
+			log.Infof("sync finished in %s", d)
+			e.clientMetrics.RecordSyncDuration(e.ctx, d)
+		})
+		e.shutdownWg.Add(1)
+		go func() {
+			defer e.shutdownWg.Done()
+			manager.run(e.ctx)
+		}()
+
+		err = e.mgmClient.Sync(e.ctx, info, manager.SetTarget)
 		if err != nil {
 			// happens if management is unavailable for a long time.
 			// We want to cancel the operation of the whole client
@@ -1315,21 +1370,104 @@ func (e *Engine) updateTURNs(turns []*mgmProto.ProtectedHostConfig) error {
 	return nil
 }
 
-func (e *Engine) updateNetworkMap(networkMap *mgmProto.NetworkMap) error {
+// updateNetworkMap applies the wholesale parts (config, routes, ACL, DNS) in full
+// and up to maxBatch peers per phase. It returns true when more peers remained
+// than the cap, so the caller re-runs until convergence.
+func (e *Engine) updateNetworkMap(networkMap *mgmProto.NetworkMap, maxBatch int, firstPass bool) (bool, error) {
 	// intentionally leave it before checking serial because for now it can happen that peer IP changed but serial didn't
 	if networkMap.GetPeerConfig() != nil {
 		err := e.updateConfig(networkMap.GetPeerConfig())
 		if err != nil {
-			return err
+			return false, err
 		}
 	}
 
 	serial := networkMap.GetSerial()
 	if e.networkSerial > serial {
 		log.Debugf("received outdated NetworkMap with serial %d, ignoring", serial)
-		return nil
+		return false, nil
 	}
 
+	// Wholesale sections (firewall/ACL, DNS, routes, forward rules) are applied
+	// up-front and only once per target: they are cheap, local, idempotent and must
+	// be in place before peers come up (fail-closed). On the bounded re-runs that only
+	// drain the remaining peer batches they are skipped — the applied forward rules are
+	// reused from e.forwardingRules for the lazy-exclude finalize.
+	if firstPass {
+		e.applyWholesale(networkMap, serial)
+	}
+
+	log.Debugf("got peers update from Management Service, total peers to connect to = %d", len(networkMap.GetRemotePeers()))
+
+	e.updateOfflinePeers(networkMap.GetOfflinePeers())
+
+	// Filter out own peer from the remote peers list
+	localPubKey := e.config.WgPrivateKey.PublicKey().String()
+	remotePeers := make([]*mgmProto.RemotePeerConfig, 0, len(networkMap.GetRemotePeers()))
+	for _, p := range networkMap.GetRemotePeers() {
+		if p.GetWgPubKey() != localPubKey {
+			remotePeers = append(remotePeers, p)
+		}
+	}
+
+	// needMore signals the caller to re-run when a peer phase hit its per-pass cap.
+	needMore := false
+
+	// cleanup request, most likely our peer has been deleted
+	if networkMap.GetRemotePeersIsEmpty() {
+		err := e.removeAllPeers()
+		e.statusRecorder.FinishPeerListModifications()
+		if err != nil {
+			return false, err
+		}
+	} else {
+		removeMore, err := e.removePeers(remotePeers, maxBatch)
+		if err != nil {
+			return false, err
+		}
+
+		modifyMore, err := e.modifyPeers(remotePeers, maxBatch)
+		if err != nil {
+			return false, err
+		}
+
+		addMore, err := e.addNewPeers(remotePeers, maxBatch)
+		if err != nil {
+			return false, err
+		}
+
+		needMore = removeMore || modifyMore || addMore
+
+		e.statusRecorder.FinishPeerListModifications()
+
+		e.updatePeerSSHHostKeys(remotePeers)
+
+		if err := e.updateSSHClientConfig(remotePeers); err != nil {
+			log.Warnf("failed to update SSH client config: %v", err)
+		}
+
+		e.updateSSHServerAuth(networkMap.GetSshAuth())
+	}
+
+	// Set the exclude list only once peers have fully converged (this pass added
+	// the last batch). It needs all target peers present in the store, and
+	// ExcludePeer has replace-semantics — a partial set mid-convergence would be wrong.
+	if !needMore {
+		excludedLazyPeers := e.toExcludedLazyPeers(e.forwardingRules, remotePeers)
+		e.connMgr.SetExcludeList(e.ctx, excludedLazyPeers)
+	}
+
+	e.networkSerial = serial
+
+	return needMore, nil
+}
+
+// applyWholesale applies the cheap, local, idempotent map sections — lazy feature
+// flag, firewall/legacy management, DNS, routes, ACL filtering, DNS forwarder and
+// ingress forward rules — that must be in place before peers come up. It runs once
+// per target (first pass only); the resulting forward rules are stashed in
+// e.forwardingRules for the lazy-exclude finalize on the peer-converged pass.
+func (e *Engine) applyWholesale(networkMap *mgmProto.NetworkMap, serial uint64) {
 	if err := e.connMgr.UpdatedRemoteFeatureFlag(e.ctx, networkMap.GetPeerConfig().GetLazyConnectionEnabled()); err != nil {
 		log.Errorf("failed to update lazy connection feature flag: %v", err)
 	}
@@ -1390,61 +1528,7 @@ func (e *Engine) updateNetworkMap(networkMap *mgmProto.NetworkMap) error {
 	if err != nil {
 		log.Errorf("failed to update forward rules, err: %v", err)
 	}
-
+	e.forwardingRules = forwardingRules
-	log.Debugf("got peers update from Management Service, total peers to connect to = %d", len(networkMap.GetRemotePeers()))
-
-	e.updateOfflinePeers(networkMap.GetOfflinePeers())
-
-	// Filter out own peer from the remote peers list
-	localPubKey := e.config.WgPrivateKey.PublicKey().String()
-	remotePeers := make([]*mgmProto.RemotePeerConfig, 0, len(networkMap.GetRemotePeers()))
-	for _, p := range networkMap.GetRemotePeers() {
-		if p.GetWgPubKey() != localPubKey {
-			remotePeers = append(remotePeers, p)
-		}
-	}
-
-	// cleanup request, most likely our peer has been deleted
-	if networkMap.GetRemotePeersIsEmpty() {
-		err := e.removeAllPeers()
-		e.statusRecorder.FinishPeerListModifications()
-		if err != nil {
-			return err
-		}
-	} else {
-		err := e.removePeers(remotePeers)
-		if err != nil {
-			return err
-		}
-
-		err = e.modifyPeers(remotePeers)
-		if err != nil {
-			return err
-		}
-
-		err = e.addNewPeers(remotePeers)
-		if err != nil {
-			return err
-		}
-
-		e.statusRecorder.FinishPeerListModifications()
-
-		e.updatePeerSSHHostKeys(remotePeers)
-
-		if err := e.updateSSHClientConfig(remotePeers); err != nil {
-			log.Warnf("failed to update SSH client config: %v", err)
-		}
-
-		e.updateSSHServerAuth(networkMap.GetSshAuth())
-	}
-
-	// must set the exclude list after the peers are added. Without it the manager can not figure out the peers parameters from the store
-	excludedLazyPeers := e.toExcludedLazyPeers(forwardingRules, remotePeers)
-	e.connMgr.SetExcludeList(e.ctx, excludedLazyPeers)
-
-	e.networkSerial = serial
-
-	return nil
 }
 
 func toDNSFeatureFlag(networkMap *mgmProto.NetworkMap) bool {
@@ -1624,14 +1708,23 @@ func addrToString(addr netip.Addr) string {
 }
 
 // addNewPeers adds peers that were not know before but arrived from the Management service with the update
-func (e *Engine) addNewPeers(peersUpdate []*mgmProto.RemotePeerConfig) error {
+// addNewPeers adds up to maxBatch not-yet-present peers per call. It returns true
+// when more new peers remained than the cap, so the caller re-runs.
+func (e *Engine) addNewPeers(peersUpdate []*mgmProto.RemotePeerConfig, maxBatch int) (bool, error) {
+	added := 0
 	for _, p := range peersUpdate {
-		err := e.addNewPeer(p)
+		if _, ok := e.peerStore.PeerConn(p.GetWgPubKey()); ok {
-		if err != nil {
+			continue // already present (cheap skip), does not count toward the cap
-			return err
 		}
+		if added >= maxBatch {
+			return true, nil // at least one more new peer remains
+		}
+		if err := e.addNewPeer(p); err != nil {
+			return false, err
+		}
+		added++
 	}
-	return nil
+	return false, nil
 }
 
 // addNewPeer add peer if connection doesn't exist

client/internal/engine_privileged_test.go

@@ -0,0 +1,565 @@
+//go:build privileged
+
+package internal
+
+import (
+	"context"
+	"fmt"
+	"net"
+	"runtime"
+	"strings"
+	"sync"
+	"testing"
+	"time"
+
+	"github.com/golang/mock/gomock"
+	"github.com/google/uuid"
+	log "github.com/sirupsen/logrus"
+	"github.com/stretchr/testify/assert"
+	"github.com/stretchr/testify/require"
+	"go.opentelemetry.io/otel"
+	"golang.zx2c4.com/wireguard/wgctrl/wgtypes"
+	"google.golang.org/grpc"
+	"google.golang.org/grpc/keepalive"
+
+	"github.com/netbirdio/netbird/client/iface"
+	"github.com/netbirdio/netbird/client/iface/device"
+	"github.com/netbirdio/netbird/client/iface/wgaddr"
+	"github.com/netbirdio/netbird/client/internal/dns"
+	"github.com/netbirdio/netbird/client/internal/peer"
+	nbssh "github.com/netbirdio/netbird/client/ssh"
+	"github.com/netbirdio/netbird/client/system"
+	nbdns "github.com/netbirdio/netbird/dns"
+	"github.com/netbirdio/netbird/management/internals/controllers/network_map/controller"
+	"github.com/netbirdio/netbird/management/internals/controllers/network_map/update_channel"
+	"github.com/netbirdio/netbird/management/internals/modules/peers"
+	"github.com/netbirdio/netbird/management/internals/modules/peers/ephemeral/manager"
+	"github.com/netbirdio/netbird/management/internals/server/config"
+	nbgrpc "github.com/netbirdio/netbird/management/internals/shared/grpc"
+	"github.com/netbirdio/netbird/management/server"
+	"github.com/netbirdio/netbird/management/server/activity"
+	nbcache "github.com/netbirdio/netbird/management/server/cache"
+	"github.com/netbirdio/netbird/management/server/groups"
+	"github.com/netbirdio/netbird/management/server/integrations/integrated_validator/validator"
+	"github.com/netbirdio/netbird/management/server/integrations/port_forwarding"
+	"github.com/netbirdio/netbird/management/server/job"
+	"github.com/netbirdio/netbird/management/server/permissions"
+	"github.com/netbirdio/netbird/management/server/settings"
+	"github.com/netbirdio/netbird/management/server/store"
+	"github.com/netbirdio/netbird/management/server/telemetry"
+	"github.com/netbirdio/netbird/management/server/types"
+	mgmt "github.com/netbirdio/netbird/shared/management/client"
+	mgmtProto "github.com/netbirdio/netbird/shared/management/proto"
+	relayClient "github.com/netbirdio/netbird/shared/relay/client"
+	signal "github.com/netbirdio/netbird/shared/signal/client"
+	"github.com/netbirdio/netbird/shared/signal/proto"
+	signalServer "github.com/netbirdio/netbird/signal/server"
+	"github.com/netbirdio/netbird/util"
+)
+
+func TestEngine_SSH(t *testing.T) {
+	key, err := wgtypes.GeneratePrivateKey()
+	if err != nil {
+		t.Fatal(err)
+		return
+	}
+
+	sshKey, err := nbssh.GeneratePrivateKey(nbssh.ED25519)
+	if err != nil {
+		t.Fatal(err)
+		return
+	}
+
+	ctx, cancel := context.WithCancel(CtxInitState(context.Background()))
+	defer cancel()
+
+	relayMgr := relayClient.NewManager(ctx, nil, key.PublicKey().String(), iface.DefaultMTU)
+	engine := NewEngine(
+		ctx, cancel,
+		&EngineConfig{
+			WgIfaceName:      "utun101",
+			WgAddr:           wgaddr.MustParseWGAddress("100.64.0.1/24"),
+			WgPrivateKey:     key,
+			WgPort:           33100,
+			ServerSSHAllowed: true,
+			MTU:              iface.DefaultMTU,
+			SSHKey:           sshKey,
+		},
+		EngineServices{
+			SignalClient:   &signal.MockClient{},
+			MgmClient:      &mgmt.MockClient{},
+			RelayManager:   relayMgr,
+			StatusRecorder: peer.NewRecorder("https://mgm"),
+		},
+		MobileDependency{},
+	)
+
+	engine.dnsServer = &dns.MockServer{
+		UpdateDNSServerFunc: func(serial uint64, update nbdns.Config) error { return nil },
+	}
+
+	err = engine.Start(nil, nil)
+	require.NoError(t, err)
+
+	defer func() {
+		err := engine.Stop()
+		if err != nil {
+			return
+		}
+	}()
+
+	peerWithSSH := &mgmtProto.RemotePeerConfig{
+		WgPubKey:   "MNHf3Ma6z6mdLbriAJbqhX7+nM/B71lgw2+91q3LfhU=",
+		AllowedIps: []string{"100.64.0.21/24"},
+		SshConfig: &mgmtProto.SSHConfig{
+			SshPubKey: []byte("ssh-ed25519 AAAAC3NzaC1lZDI1NTE5AAAAIFATYCqaQw/9id1Qkq3n16JYhDhXraI6Pc1fgB8ynEfQ"),
+		},
+	}
+
+	// SSH server is not enabled so SSH config of a remote peer should be ignored
+	networkMap := &mgmtProto.NetworkMap{
+		Serial:             6,
+		PeerConfig:         nil,
+		RemotePeers:        []*mgmtProto.RemotePeerConfig{peerWithSSH},
+		RemotePeersIsEmpty: false,
+	}
+
+	_, err = engine.updateNetworkMap(networkMap, maxPeersPerSyncPass, true)
+	require.NoError(t, err)
+
+	assert.Nil(t, engine.sshServer)
+
+	// SSH server is enabled, therefore SSH config should be applied
+	networkMap = &mgmtProto.NetworkMap{
+		Serial: 7,
+		PeerConfig: &mgmtProto.PeerConfig{Address: "100.64.0.1/24",
+			SshConfig: &mgmtProto.SSHConfig{
+				SshEnabled: true,
+				JwtConfig: &mgmtProto.JWTConfig{
+					Issuer:       "test-issuer",
+					Audience:     "test-audience",
+					KeysLocation: "test-keys",
+					MaxTokenAge:  3600,
+				},
+			}},
+		RemotePeers:        []*mgmtProto.RemotePeerConfig{peerWithSSH},
+		RemotePeersIsEmpty: false,
+	}
+
+	_, err = engine.updateNetworkMap(networkMap, maxPeersPerSyncPass, true)
+	require.NoError(t, err)
+
+	time.Sleep(250 * time.Millisecond)
+	assert.NotNil(t, engine.sshServer)
+
+	// now remove peer
+	networkMap = &mgmtProto.NetworkMap{
+		Serial:             8,
+		RemotePeers:        []*mgmtProto.RemotePeerConfig{},
+		RemotePeersIsEmpty: false,
+	}
+
+	_, err = engine.updateNetworkMap(networkMap, maxPeersPerSyncPass, true)
+	require.NoError(t, err)
+
+	// time.Sleep(250 * time.Millisecond)
+	assert.NotNil(t, engine.sshServer)
+
+	// now disable SSH server
+	networkMap = &mgmtProto.NetworkMap{
+		Serial: 9,
+		PeerConfig: &mgmtProto.PeerConfig{Address: "100.64.0.1/24",
+			SshConfig: &mgmtProto.SSHConfig{SshEnabled: false}},
+		RemotePeers:        []*mgmtProto.RemotePeerConfig{peerWithSSH},
+		RemotePeersIsEmpty: false,
+	}
+
+	_, err = engine.updateNetworkMap(networkMap, maxPeersPerSyncPass, true)
+	require.NoError(t, err)
+
+	assert.Nil(t, engine.sshServer)
+}
+
+func TestEngine_Sync(t *testing.T) {
+	key, err := wgtypes.GeneratePrivateKey()
+	if err != nil {
+		t.Fatal(err)
+		return
+	}
+
+	ctx, cancel := context.WithCancel(CtxInitState(context.Background()))
+	defer cancel()
+
+	// feed updates to Engine via mocked Management client
+	updates := make(chan *mgmtProto.SyncResponse)
+	defer close(updates)
+	syncFunc := func(ctx context.Context, info *system.Info, msgHandler func(msg *mgmtProto.SyncResponse) error) error {
+		for msg := range updates {
+			err := msgHandler(msg)
+			if err != nil {
+				t.Fatal(err)
+			}
+		}
+		return nil
+	}
+	relayMgr := relayClient.NewManager(ctx, nil, key.PublicKey().String(), iface.DefaultMTU)
+	engine := NewEngine(ctx, cancel, &EngineConfig{
+		WgIfaceName:  "utun103",
+		WgAddr:       wgaddr.MustParseWGAddress("100.64.0.1/24"),
+		WgPrivateKey: key,
+		WgPort:       33100,
+		MTU:          iface.DefaultMTU,
+	}, EngineServices{
+		SignalClient:   &signal.MockClient{},
+		MgmClient:      &mgmt.MockClient{SyncFunc: syncFunc},
+		RelayManager:   relayMgr,
+		StatusRecorder: peer.NewRecorder("https://mgm"),
+	}, MobileDependency{})
+	engine.ctx = ctx
+
+	engine.dnsServer = &dns.MockServer{
+		UpdateDNSServerFunc: func(serial uint64, update nbdns.Config) error { return nil },
+	}
+
+	defer func() {
+		err := engine.Stop()
+		if err != nil {
+			return
+		}
+	}()
+
+	err = engine.Start(nil, nil)
+	if err != nil {
+		t.Fatal(err)
+		return
+	}
+
+	peer1 := &mgmtProto.RemotePeerConfig{
+		WgPubKey:   "RRHf3Ma6z6mdLbriAJbqhX7+nM/B71lgw2+91q3LfhU=",
+		AllowedIps: []string{"100.64.0.10/24"},
+	}
+	peer2 := &mgmtProto.RemotePeerConfig{
+		WgPubKey:   "LLHf3Ma6z6mdLbriAJbqhX9+nM/B71lgw2+91q3LlhU=",
+		AllowedIps: []string{"100.64.0.11/24"},
+	}
+	peer3 := &mgmtProto.RemotePeerConfig{
+		WgPubKey:   "GGHf3Ma6z6mdLbriAJbqhX9+nM/B71lgw2+91q3LlhU=",
+		AllowedIps: []string{"100.64.0.12/24"},
+	}
+	// 1st update with just 1 peer and serial larger than the current serial of the engine => apply update
+	updates <- &mgmtProto.SyncResponse{
+		NetworkMap: &mgmtProto.NetworkMap{
+			Serial:             10,
+			PeerConfig:         nil,
+			RemotePeers:        []*mgmtProto.RemotePeerConfig{peer1, peer2, peer3},
+			RemotePeersIsEmpty: false,
+		},
+	}
+
+	timeout := time.After(time.Second * 2)
+	for {
+		select {
+		case <-timeout:
+			t.Fatalf("timeout while waiting for test to finish")
+			return
+		default:
+		}
+
+		if getPeers(engine) == 3 && engine.networkSerial == 10 {
+			break
+		}
+	}
+}
+
+func TestEngine_MultiplePeers(t *testing.T) {
+	// log.SetLevel(log.DebugLevel)
+
+	ctx, cancel := context.WithCancel(CtxInitState(context.Background()))
+	defer cancel()
+
+	sigServer, signalAddr, err := startSignal(t)
+	if err != nil {
+		t.Fatal(err)
+		return
+	}
+	defer sigServer.Stop()
+	mgmtServer, mgmtAddr, err := startManagement(t, t.TempDir(), "../testdata/store.sql")
+	if err != nil {
+		t.Fatal(err)
+		return
+	}
+	defer mgmtServer.GracefulStop()
+
+	setupKey := "A2C8E62B-38F5-4553-B31E-DD66C696CEBB"
+
+	mu := sync.Mutex{}
+	engines := []*Engine{}
+	numPeers := 10
+	wg := sync.WaitGroup{}
+	wg.Add(numPeers)
+	// create and start peers
+	for i := 0; i < numPeers; i++ {
+		j := i
+		go func() {
+			engine, err := createEngine(ctx, cancel, setupKey, j, mgmtAddr, signalAddr)
+			if err != nil {
+				wg.Done()
+				t.Errorf("unable to create the engine for peer %d with error %v", j, err)
+				return
+			}
+			engine.dnsServer = &dns.MockServer{}
+			mu.Lock()
+			defer mu.Unlock()
+			guid := fmt.Sprintf("{%s}", uuid.New().String())
+			device.CustomWindowsGUIDString = strings.ToLower(guid)
+			err = engine.Start(nil, nil)
+			if err != nil {
+				t.Errorf("unable to start engine for peer %d with error %v", j, err)
+				wg.Done()
+				return
+			}
+			engines = append(engines, engine)
+			wg.Done()
+		}()
+	}
+
+	// wait until all have been created and started
+	wg.Wait()
+	if len(engines) != numPeers {
+		t.Fatal("not all peers were started")
+	}
+	// check whether all the peer have expected peers connected
+
+	expectedConnected := numPeers * (numPeers - 1)
+
+	// adjust according to timeouts
+	timeout := 50 * time.Second
+	timeoutChan := time.After(timeout)
+	ticker := time.NewTicker(time.Second)
+	defer ticker.Stop()
+loop:
+	for {
+		select {
+		case <-timeoutChan:
+			t.Fatalf("waiting for expected connections timeout after %s", timeout.String())
+			break loop
+		case <-ticker.C:
+			totalConnected := 0
+			for _, engine := range engines {
+				totalConnected += getConnectedPeers(engine)
+			}
+			if totalConnected == expectedConnected {
+				log.Infof("total connected=%d", totalConnected)
+				break loop
+			}
+			log.Infof("total connected=%d", totalConnected)
+		}
+	}
+	// cleanup test
+	for n, peerEngine := range engines {
+		t.Logf("stopping peer with interface %s from multipeer test, loopIndex %d", peerEngine.wgInterface.Name(), n)
+		errStop := peerEngine.mgmClient.Close()
+		if errStop != nil {
+			log.Infoln("got error trying to close management clients from engine: ", errStop)
+		}
+		errStop = peerEngine.Stop()
+		if errStop != nil {
+			log.Infoln("got error trying to close testing peers engine: ", errStop)
+		}
+	}
+}
+
+var (
+	kaep = keepalive.EnforcementPolicy{
+		MinTime:             15 * time.Second,
+		PermitWithoutStream: true,
+	}
+
+	kasp = keepalive.ServerParameters{
+		MaxConnectionIdle:     15 * time.Second,
+		MaxConnectionAgeGrace: 5 * time.Second,
+		Time:                  5 * time.Second,
+		Timeout:               2 * time.Second,
+	}
+)
+
+func createEngine(ctx context.Context, cancel context.CancelFunc, setupKey string, i int, mgmtAddr string, signalAddr string) (*Engine, error) {
+	key, err := wgtypes.GeneratePrivateKey()
+	if err != nil {
+		return nil, err
+	}
+	mgmtClient, err := mgmt.NewClient(ctx, mgmtAddr, key, false)
+	if err != nil {
+		return nil, err
+	}
+	signalClient, err := signal.NewClient(ctx, signalAddr, key, false)
+	if err != nil {
+		return nil, err
+	}
+
+	info := system.GetInfo(ctx)
+	resp, err := mgmtClient.Register(setupKey, "", info, nil, nil)
+	if err != nil {
+		return nil, err
+	}
+
+	var ifaceName string
+	if runtime.GOOS == "darwin" {
+		ifaceName = fmt.Sprintf("utun1%d", i)
+	} else {
+		ifaceName = fmt.Sprintf("wt%d", i)
+	}
+
+	wgPort := 33100 + i
+	conf := &EngineConfig{
+		WgIfaceName:  ifaceName,
+		WgAddr:       wgaddr.MustParseWGAddress(resp.PeerConfig.Address),
+		WgPrivateKey: key,
+		WgPort:       wgPort,
+		MTU:          iface.DefaultMTU,
+	}
+
+	relayMgr := relayClient.NewManager(ctx, nil, key.PublicKey().String(), iface.DefaultMTU)
+	e, err := NewEngine(ctx, cancel, conf, EngineServices{
+		SignalClient:   signalClient,
+		MgmClient:      mgmtClient,
+		RelayManager:   relayMgr,
+		StatusRecorder: peer.NewRecorder("https://mgm"),
+	}, MobileDependency{}), nil
+	e.ctx = ctx
+	return e, err
+}
+
+func startSignal(t *testing.T) (*grpc.Server, string, error) {
+	t.Helper()
+
+	s := grpc.NewServer(grpc.KeepaliveEnforcementPolicy(kaep), grpc.KeepaliveParams(kasp))
+
+	lis, err := net.Listen("tcp", "localhost:0")
+	if err != nil {
+		log.Fatalf("failed to listen: %v", err)
+	}
+
+	srv, err := signalServer.NewServer(context.Background(), otel.Meter(""))
+	require.NoError(t, err)
+	proto.RegisterSignalExchangeServer(s, srv)
+
+	go func() {
+		if err = s.Serve(lis); err != nil {
+			log.Fatalf("failed to serve: %v", err)
+		}
+	}()
+
+	return s, lis.Addr().String(), nil
+}
+
+func startManagement(t *testing.T, dataDir, testFile string) (*grpc.Server, string, error) {
+	t.Helper()
+
+	config := &config.Config{
+		Stuns:      []*config.Host{},
+		TURNConfig: &config.TURNConfig{},
+		Relay: &config.Relay{
+			Addresses:      []string{"127.0.0.1:1234"},
+			CredentialsTTL: util.Duration{Duration: time.Hour},
+			Secret:         "222222222222222222",
+		},
+		Signal: &config.Host{
+			Proto: "http",
+			URI:   "localhost:10000",
+		},
+		Datadir:    dataDir,
+		HttpConfig: nil,
+	}
+
+	lis, err := net.Listen("tcp", "localhost:0")
+	if err != nil {
+		return nil, "", err
+	}
+	s := grpc.NewServer(grpc.KeepaliveEnforcementPolicy(kaep), grpc.KeepaliveParams(kasp))
+
+	store, cleanUp, err := store.NewTestStoreFromSQL(context.Background(), testFile, config.Datadir)
+	if err != nil {
+		return nil, "", err
+	}
+	t.Cleanup(cleanUp)
+
+	eventStore := &activity.InMemoryEventStore{}
+	if err != nil {
+		return nil, "", err
+	}
+
+	permissionsManager := permissions.NewManager(store)
+	peersManager := peers.NewManager(store, permissionsManager)
+	jobManager := job.NewJobManager(nil, store, peersManager)
+
+	cacheStore, err := nbcache.NewStore(context.Background(), 100*time.Millisecond, 300*time.Millisecond, 100)
+	if err != nil {
+		return nil, "", err
+	}
+
+	ia, _ := validator.NewIntegratedValidator(context.Background(), peersManager, nil, eventStore, cacheStore)
+
+	metrics, err := telemetry.NewDefaultAppMetrics(context.Background())
+	require.NoError(t, err)
+
+	ctrl := gomock.NewController(t)
+	t.Cleanup(ctrl.Finish)
+	settingsMockManager := settings.NewMockManager(ctrl)
+	settingsMockManager.EXPECT().
+		GetSettings(gomock.Any(), gomock.Any(), gomock.Any()).
+		Return(&types.Settings{}, nil).
+		AnyTimes()
+	settingsMockManager.EXPECT().
+		GetExtraSettings(gomock.Any(), gomock.Any()).
+		Return(&types.ExtraSettings{}, nil).
+		AnyTimes()
+
+	groupsManager := groups.NewManagerMock()
+
+	updateManager := update_channel.NewPeersUpdateManager(metrics)
+	requestBuffer := server.NewAccountRequestBuffer(context.Background(), store)
+	networkMapController := controller.NewController(context.Background(), store, metrics, updateManager, requestBuffer, server.MockIntegratedValidator{}, settingsMockManager, "netbird.selfhosted", port_forwarding.NewControllerMock(), manager.NewEphemeralManager(store, peersManager), config)
+	accountManager, err := server.BuildManager(context.Background(), config, store, networkMapController, jobManager, nil, "", eventStore, nil, false, ia, metrics, port_forwarding.NewControllerMock(), settingsMockManager, permissionsManager, false, cacheStore)
+	if err != nil {
+		return nil, "", err
+	}
+
+	secretsManager, err := nbgrpc.NewTimeBasedAuthSecretsManager(updateManager, config.TURNConfig, config.Relay, settingsMockManager, groupsManager)
+	if err != nil {
+		return nil, "", err
+	}
+	mgmtServer, err := nbgrpc.NewServer(config, accountManager, settingsMockManager, jobManager, secretsManager, nil, nil, &server.MockIntegratedValidator{}, networkMapController, nil, nil)
+	if err != nil {
+		return nil, "", err
+	}
+	mgmtProto.RegisterManagementServiceServer(s, mgmtServer)
+	go func() {
+		if err = s.Serve(lis); err != nil {
+			log.Fatalf("failed to serve: %v", err)
+		}
+	}()
+
+	return s, lis.Addr().String(), nil
+}
+
+// getConnectedPeers returns a connection Status or nil if peer connection wasn't found
+func getConnectedPeers(e *Engine) int {
+	e.syncMsgMux.Lock()
+	defer e.syncMsgMux.Unlock()
+	i := 0
+	for _, id := range e.peerStore.PeersPubKey() {
+		conn, _ := e.peerStore.PeerConn(id)
+		if conn.IsConnected() {
+			i++
+		}
+	}
+	return i
+}
+
+func getPeers(e *Engine) int {
+	e.syncMsgMux.Lock()
+	defer e.syncMsgMux.Unlock()
+
+	return len(e.peerStore.PeersPubKey())
+}

client/internal/engine_test.go

@@ -6,37 +6,18 @@ import (
 	"net"
 	"net/netip"
 	"os"
-	"runtime"
 	"strings"
 	"sync"
 	"testing"
 	"time"
 
-	"github.com/golang/mock/gomock"
-	"github.com/google/uuid"
-	log "github.com/sirupsen/logrus"
 	"github.com/stretchr/testify/assert"
 	"github.com/stretchr/testify/require"
-	"go.opentelemetry.io/otel"
 	wgdevice "golang.zx2c4.com/wireguard/device"
 	"golang.zx2c4.com/wireguard/tun/netstack"
 	"golang.zx2c4.com/wireguard/wgctrl/wgtypes"
-	"google.golang.org/grpc"
-	"google.golang.org/grpc/keepalive"
 
 	"github.com/netbirdio/netbird/client/internal/stdnet"
-	"github.com/netbirdio/netbird/management/server/job"
-
-	"github.com/netbirdio/netbird/management/server/integrations/integrated_validator/validator"
-
-	"github.com/netbirdio/netbird/management/internals/controllers/network_map/controller"
-	"github.com/netbirdio/netbird/management/internals/controllers/network_map/update_channel"
-	"github.com/netbirdio/netbird/management/internals/modules/peers"
-	"github.com/netbirdio/netbird/management/internals/modules/peers/ephemeral/manager"
-	nbgrpc "github.com/netbirdio/netbird/management/internals/shared/grpc"
-
-	"github.com/netbirdio/netbird/management/internals/server/config"
-	"github.com/netbirdio/netbird/management/server/groups"
 
 	"github.com/netbirdio/netbird/client/iface"
 	"github.com/netbirdio/netbird/client/iface/configurer"
@@ -50,18 +31,7 @@ import (
 	icemaker "github.com/netbirdio/netbird/client/internal/peer/ice"
 	"github.com/netbirdio/netbird/client/internal/profilemanager"
 	"github.com/netbirdio/netbird/client/internal/routemanager"
-	nbssh "github.com/netbirdio/netbird/client/ssh"
-	"github.com/netbirdio/netbird/client/system"
 	nbdns "github.com/netbirdio/netbird/dns"
-	"github.com/netbirdio/netbird/management/server"
-	"github.com/netbirdio/netbird/management/server/activity"
-	nbcache "github.com/netbirdio/netbird/management/server/cache"
-	"github.com/netbirdio/netbird/management/server/integrations/port_forwarding"
-	"github.com/netbirdio/netbird/management/server/permissions"
-	"github.com/netbirdio/netbird/management/server/settings"
-	"github.com/netbirdio/netbird/management/server/store"
-	"github.com/netbirdio/netbird/management/server/telemetry"
-	"github.com/netbirdio/netbird/management/server/types"
 	"github.com/netbirdio/netbird/monotime"
 	"github.com/netbirdio/netbird/route"
 	mgmt "github.com/netbirdio/netbird/shared/management/client"
@@ -69,25 +39,9 @@ import (
 	"github.com/netbirdio/netbird/shared/netiputil"
 	relayClient "github.com/netbirdio/netbird/shared/relay/client"
 	signal "github.com/netbirdio/netbird/shared/signal/client"
-	"github.com/netbirdio/netbird/shared/signal/proto"
-	signalServer "github.com/netbirdio/netbird/signal/server"
 	"github.com/netbirdio/netbird/util"
 )
 
-var (
-	kaep = keepalive.EnforcementPolicy{
-		MinTime:             15 * time.Second,
-		PermitWithoutStream: true,
-	}
-
-	kasp = keepalive.ServerParameters{
-		MaxConnectionIdle:     15 * time.Second,
-		MaxConnectionAgeGrace: 5 * time.Second,
-		Time:                  5 * time.Second,
-		Timeout:               2 * time.Second,
-	}
-)
-
 type MockWGIface struct {
 	CreateFunc                 func() error
 	CreateOnAndroidFunc        func(routeRange []string, ip string, domains []string) error
@@ -234,129 +188,6 @@ func TestMain(m *testing.M) {
 	os.Exit(code)
 }
 
-func TestEngine_SSH(t *testing.T) {
-	key, err := wgtypes.GeneratePrivateKey()
-	if err != nil {
-		t.Fatal(err)
-		return
-	}
-
-	sshKey, err := nbssh.GeneratePrivateKey(nbssh.ED25519)
-	if err != nil {
-		t.Fatal(err)
-		return
-	}
-
-	ctx, cancel := context.WithCancel(CtxInitState(context.Background()))
-	defer cancel()
-
-	relayMgr := relayClient.NewManager(ctx, nil, key.PublicKey().String(), iface.DefaultMTU)
-	engine := NewEngine(
-		ctx, cancel,
-		&EngineConfig{
-			WgIfaceName:      "utun101",
-			WgAddr:           wgaddr.MustParseWGAddress("100.64.0.1/24"),
-			WgPrivateKey:     key,
-			WgPort:           33100,
-			ServerSSHAllowed: true,
-			MTU:              iface.DefaultMTU,
-			SSHKey:           sshKey,
-		},
-		EngineServices{
-			SignalClient:   &signal.MockClient{},
-			MgmClient:      &mgmt.MockClient{},
-			RelayManager:   relayMgr,
-			StatusRecorder: peer.NewRecorder("https://mgm"),
-		},
-		MobileDependency{},
-	)
-
-	engine.dnsServer = &dns.MockServer{
-		UpdateDNSServerFunc: func(serial uint64, update nbdns.Config) error { return nil },
-	}
-
-	err = engine.Start(nil, nil)
-	require.NoError(t, err)
-
-	defer func() {
-		err := engine.Stop()
-		if err != nil {
-			return
-		}
-	}()
-
-	peerWithSSH := &mgmtProto.RemotePeerConfig{
-		WgPubKey:   "MNHf3Ma6z6mdLbriAJbqhX7+nM/B71lgw2+91q3LfhU=",
-		AllowedIps: []string{"100.64.0.21/24"},
-		SshConfig: &mgmtProto.SSHConfig{
-			SshPubKey: []byte("ssh-ed25519 AAAAC3NzaC1lZDI1NTE5AAAAIFATYCqaQw/9id1Qkq3n16JYhDhXraI6Pc1fgB8ynEfQ"),
-		},
-	}
-
-	// SSH server is not enabled so SSH config of a remote peer should be ignored
-	networkMap := &mgmtProto.NetworkMap{
-		Serial:             6,
-		PeerConfig:         nil,
-		RemotePeers:        []*mgmtProto.RemotePeerConfig{peerWithSSH},
-		RemotePeersIsEmpty: false,
-	}
-
-	err = engine.updateNetworkMap(networkMap)
-	require.NoError(t, err)
-
-	assert.Nil(t, engine.sshServer)
-
-	// SSH server is enabled, therefore SSH config should be applied
-	networkMap = &mgmtProto.NetworkMap{
-		Serial: 7,
-		PeerConfig: &mgmtProto.PeerConfig{Address: "100.64.0.1/24",
-			SshConfig: &mgmtProto.SSHConfig{
-				SshEnabled: true,
-				JwtConfig: &mgmtProto.JWTConfig{
-					Issuer:       "test-issuer",
-					Audience:     "test-audience",
-					KeysLocation: "test-keys",
-					MaxTokenAge:  3600,
-				},
-			}},
-		RemotePeers:        []*mgmtProto.RemotePeerConfig{peerWithSSH},
-		RemotePeersIsEmpty: false,
-	}
-
-	err = engine.updateNetworkMap(networkMap)
-	require.NoError(t, err)
-
-	time.Sleep(250 * time.Millisecond)
-	assert.NotNil(t, engine.sshServer)
-
-	// now remove peer
-	networkMap = &mgmtProto.NetworkMap{
-		Serial:             8,
-		RemotePeers:        []*mgmtProto.RemotePeerConfig{},
-		RemotePeersIsEmpty: false,
-	}
-
-	err = engine.updateNetworkMap(networkMap)
-	require.NoError(t, err)
-
-	// time.Sleep(250 * time.Millisecond)
-	assert.NotNil(t, engine.sshServer)
-
-	// now disable SSH server
-	networkMap = &mgmtProto.NetworkMap{
-		Serial: 9,
-		PeerConfig: &mgmtProto.PeerConfig{Address: "100.64.0.1/24",
-			SshConfig: &mgmtProto.SSHConfig{SshEnabled: false}},
-		RemotePeers:        []*mgmtProto.RemotePeerConfig{peerWithSSH},
-		RemotePeersIsEmpty: false,
-	}
-
-	err = engine.updateNetworkMap(networkMap)
-	require.NoError(t, err)
-
-	assert.Nil(t, engine.sshServer)
-}
-
 func TestEngine_SSHUpdateLogic(t *testing.T) {
 	// Test that SSH server start/stop logic works based on config
 	engine := &Engine{
@@ -602,7 +433,7 @@ func TestEngine_UpdateNetworkMap(t *testing.T) {
 
 	for _, c := range []testCase{case1, case2, case3, case4, case5, case6} {
 		t.Run(c.name, func(t *testing.T) {
-			err = engine.updateNetworkMap(c.networkMap)
+			_, err = engine.updateNetworkMap(c.networkMap, maxPeersPerSyncPass, true)
 			if err != nil {
 				t.Fatal(err)
 				return
@@ -629,97 +460,47 @@ func TestEngine_UpdateNetworkMap(t *testing.T) {
 			}
 		})
 	}
-}
 
-func TestEngine_Sync(t *testing.T) {
+	// chunked apply: with a per-pass cap smaller than the number of peers, a
-	key, err := wgtypes.GeneratePrivateKey()
+	// single updateNetworkMap applies one batch and reports more==true; the
-	if err != nil {
+	// caller re-runs until convergence. (engine currently holds 0 peers.)
-		t.Fatal(err)
+	t.Run("chunked add converges over multiple passes", func(t *testing.T) {
-		return
+		nm := &mgmtProto.NetworkMap{
-	}
+			Serial:      6,
-
+			RemotePeers: []*mgmtProto.RemotePeerConfig{peer1, peer2, peer3},
-	ctx, cancel := context.WithCancel(CtxInitState(context.Background()))
-	defer cancel()
-
-	// feed updates to Engine via mocked Management client
-	updates := make(chan *mgmtProto.SyncResponse)
-	defer close(updates)
-	syncFunc := func(ctx context.Context, info *system.Info, msgHandler func(msg *mgmtProto.SyncResponse) error) error {
-		for msg := range updates {
-			err := msgHandler(msg)
-			if err != nil {
-				t.Fatal(err)
-			}
-		}
-		return nil
-	}
-	relayMgr := relayClient.NewManager(ctx, nil, key.PublicKey().String(), iface.DefaultMTU)
-	engine := NewEngine(ctx, cancel, &EngineConfig{
-		WgIfaceName:  "utun103",
-		WgAddr:       wgaddr.MustParseWGAddress("100.64.0.1/24"),
-		WgPrivateKey: key,
-		WgPort:       33100,
-		MTU:          iface.DefaultMTU,
-	}, EngineServices{
-		SignalClient:   &signal.MockClient{},
-		MgmClient:      &mgmt.MockClient{SyncFunc: syncFunc},
-		RelayManager:   relayMgr,
-		StatusRecorder: peer.NewRecorder("https://mgm"),
-	}, MobileDependency{})
-	engine.ctx = ctx
-
-	engine.dnsServer = &dns.MockServer{
-		UpdateDNSServerFunc: func(serial uint64, update nbdns.Config) error { return nil },
-	}
-
-	defer func() {
-		err := engine.Stop()
-		if err != nil {
-			return
-		}
-	}()
-
-	err = engine.Start(nil, nil)
-	if err != nil {
-		t.Fatal(err)
-		return
-	}
-
-	peer1 := &mgmtProto.RemotePeerConfig{
-		WgPubKey:   "RRHf3Ma6z6mdLbriAJbqhX7+nM/B71lgw2+91q3LfhU=",
-		AllowedIps: []string{"100.64.0.10/24"},
-	}
-	peer2 := &mgmtProto.RemotePeerConfig{
-		WgPubKey:   "LLHf3Ma6z6mdLbriAJbqhX9+nM/B71lgw2+91q3LlhU=",
-		AllowedIps: []string{"100.64.0.11/24"},
-	}
-	peer3 := &mgmtProto.RemotePeerConfig{
-		WgPubKey:   "GGHf3Ma6z6mdLbriAJbqhX9+nM/B71lgw2+91q3LlhU=",
-		AllowedIps: []string{"100.64.0.12/24"},
-	}
-	// 1st update with just 1 peer and serial larger than the current serial of the engine => apply update
-	updates <- &mgmtProto.SyncResponse{
-		NetworkMap: &mgmtProto.NetworkMap{
-			Serial:             10,
-			PeerConfig:         nil,
-			RemotePeers:        []*mgmtProto.RemotePeerConfig{peer1, peer2, peer3},
-			RemotePeersIsEmpty: false,
-		},
-	}
-
-	timeout := time.After(time.Second * 2)
-	for {
-		select {
-		case <-timeout:
-			t.Fatalf("timeout while waiting for test to finish")
-			return
-		default:
 		}
 
-		if getPeers(engine) == 3 && engine.networkSerial == 10 {
+		more, err := engine.updateNetworkMap(nm, 1, true)
-			break
+		require.NoError(t, err)
+		require.True(t, more, "pass 1 should signal more")
+		require.Len(t, engine.peerStore.PeersPubKey(), 1)
+
+		more, err = engine.updateNetworkMap(nm, 1, false)
+		require.NoError(t, err)
+		require.True(t, more, "pass 2 should signal more")
+		require.Len(t, engine.peerStore.PeersPubKey(), 2)
+
+		more, err = engine.updateNetworkMap(nm, 1, false)
+		require.NoError(t, err)
+		require.False(t, more, "pass 3 should converge")
+		require.Len(t, engine.peerStore.PeersPubKey(), 3)
+	})
+
+	t.Run("chunked remove converges over multiple passes", func(t *testing.T) {
+		nm := &mgmtProto.NetworkMap{
+			Serial:      7,
+			RemotePeers: []*mgmtProto.RemotePeerConfig{peer1}, // remove peer2, peer3
 		}
-	}
+
+		more, err := engine.updateNetworkMap(nm, 1, true)
+		require.NoError(t, err)
+		require.True(t, more, "pass 1 should signal more (2 to remove, cap 1)")
+
+		more, err = engine.updateNetworkMap(nm, 1, false)
+		require.NoError(t, err)
+		require.False(t, more, "pass 2 should converge")
+		require.Len(t, engine.peerStore.PeersPubKey(), 1)
+	})
 }
 
 func TestEngine_UpdateNetworkMapWithRoutes(t *testing.T) {
@@ -890,7 +671,7 @@ func TestEngine_UpdateNetworkMapWithRoutes(t *testing.T) {
 				}
 			}()
 
-			err = engine.updateNetworkMap(testCase.networkMap)
+			_, err = engine.updateNetworkMap(testCase.networkMap, maxPeersPerSyncPass, true)
 			assert.NoError(t, err, "shouldn't return error")
 			assert.Equal(t, testCase.expectedSerial, input.inputSerial, "serial should match")
 			assert.Len(t, input.clientRoutes, testCase.expectedLen, "clientRoutes len should match")
@@ -1094,7 +875,7 @@ func TestEngine_UpdateNetworkMapWithDNSUpdate(t *testing.T) {
 				}
 			}()
 
-			err = engine.updateNetworkMap(testCase.networkMap)
+			_, err = engine.updateNetworkMap(testCase.networkMap, maxPeersPerSyncPass, true)
 			assert.NoError(t, err, "shouldn't return error")
 			assert.Equal(t, testCase.expectedSerial, input.inputSerial, "serial should match")
 			assert.Len(t, input.inputNSGroups, testCase.expectedZonesLen, "zones len should match")
@@ -1105,104 +886,6 @@ func TestEngine_UpdateNetworkMapWithDNSUpdate(t *testing.T) {
 	}
 }
 
-func TestEngine_MultiplePeers(t *testing.T) {
-	// log.SetLevel(log.DebugLevel)
-
-	ctx, cancel := context.WithCancel(CtxInitState(context.Background()))
-	defer cancel()
-
-	sigServer, signalAddr, err := startSignal(t)
-	if err != nil {
-		t.Fatal(err)
-		return
-	}
-	defer sigServer.Stop()
-	mgmtServer, mgmtAddr, err := startManagement(t, t.TempDir(), "../testdata/store.sql")
-	if err != nil {
-		t.Fatal(err)
-		return
-	}
-	defer mgmtServer.GracefulStop()
-
-	setupKey := "A2C8E62B-38F5-4553-B31E-DD66C696CEBB"
-
-	mu := sync.Mutex{}
-	engines := []*Engine{}
-	numPeers := 10
-	wg := sync.WaitGroup{}
-	wg.Add(numPeers)
-	// create and start peers
-	for i := 0; i < numPeers; i++ {
-		j := i
-		go func() {
-			engine, err := createEngine(ctx, cancel, setupKey, j, mgmtAddr, signalAddr)
-			if err != nil {
-				wg.Done()
-				t.Errorf("unable to create the engine for peer %d with error %v", j, err)
-				return
-			}
-			engine.dnsServer = &dns.MockServer{}
-			mu.Lock()
-			defer mu.Unlock()
-			guid := fmt.Sprintf("{%s}", uuid.New().String())
-			device.CustomWindowsGUIDString = strings.ToLower(guid)
-			err = engine.Start(nil, nil)
-			if err != nil {
-				t.Errorf("unable to start engine for peer %d with error %v", j, err)
-				wg.Done()
-				return
-			}
-			engines = append(engines, engine)
-			wg.Done()
-		}()
-	}
-
-	// wait until all have been created and started
-	wg.Wait()
-	if len(engines) != numPeers {
-		t.Fatal("not all peers was started")
-	}
-	// check whether all the peer have expected peers connected
-
-	expectedConnected := numPeers * (numPeers - 1)
-
-	// adjust according to timeouts
-	timeout := 50 * time.Second
-	timeoutChan := time.After(timeout)
-	ticker := time.NewTicker(time.Second)
-	defer ticker.Stop()
-loop:
-	for {
-		select {
-		case <-timeoutChan:
-			t.Fatalf("waiting for expected connections timeout after %s", timeout.String())
-			break loop
-		case <-ticker.C:
-			totalConnected := 0
-			for _, engine := range engines {
-				totalConnected += getConnectedPeers(engine)
-			}
-			if totalConnected == expectedConnected {
-				log.Infof("total connected=%d", totalConnected)
-				break loop
-			}
-			log.Infof("total connected=%d", totalConnected)
-		}
-	}
-	// cleanup test
-	for n, peerEngine := range engines {
-		t.Logf("stopping peer with interface %s from multipeer test, loopIndex %d", peerEngine.wgInterface.Name(), n)
-		errStop := peerEngine.mgmClient.Close()
-		if errStop != nil {
-			log.Infoln("got error trying to close management clients from engine: ", errStop)
-		}
-		errStop = peerEngine.Stop()
-		if errStop != nil {
-			log.Infoln("got error trying to close testing peers engine: ", errStop)
-		}
-	}
-}
-
 func Test_ParseNATExternalIPMappings(t *testing.T) {
 	ifaceList, err := net.Interfaces()
 	if err != nil {
@@ -1526,187 +1209,6 @@ func TestCompareNetIPLists(t *testing.T) {
 	}
 }
 
-func createEngine(ctx context.Context, cancel context.CancelFunc, setupKey string, i int, mgmtAddr string, signalAddr string) (*Engine, error) {
-	key, err := wgtypes.GeneratePrivateKey()
-	if err != nil {
-		return nil, err
-	}
-	mgmtClient, err := mgmt.NewClient(ctx, mgmtAddr, key, false)
-	if err != nil {
-		return nil, err
-	}
-	signalClient, err := signal.NewClient(ctx, signalAddr, key, false)
-	if err != nil {
-		return nil, err
-	}
-
-	info := system.GetInfo(ctx)
-	resp, err := mgmtClient.Register(setupKey, "", info, nil, nil)
-	if err != nil {
-		return nil, err
-	}
-
-	var ifaceName string
-	if runtime.GOOS == "darwin" {
-		ifaceName = fmt.Sprintf("utun1%d", i)
-	} else {
-		ifaceName = fmt.Sprintf("wt%d", i)
-	}
-
-	wgPort := 33100 + i
-	conf := &EngineConfig{
-		WgIfaceName:  ifaceName,
-		WgAddr:       wgaddr.MustParseWGAddress(resp.PeerConfig.Address),
-		WgPrivateKey: key,
-		WgPort:       wgPort,
-		MTU:          iface.DefaultMTU,
-	}
-
-	relayMgr := relayClient.NewManager(ctx, nil, key.PublicKey().String(), iface.DefaultMTU)
-	e, err := NewEngine(ctx, cancel, conf, EngineServices{
-		SignalClient:   signalClient,
-		MgmClient:      mgmtClient,
-		RelayManager:   relayMgr,
-		StatusRecorder: peer.NewRecorder("https://mgm"),
-	}, MobileDependency{}), nil
-	e.ctx = ctx
-	return e, err
-}
-
-func startSignal(t *testing.T) (*grpc.Server, string, error) {
-	t.Helper()
-
-	s := grpc.NewServer(grpc.KeepaliveEnforcementPolicy(kaep), grpc.KeepaliveParams(kasp))
-
-	lis, err := net.Listen("tcp", "localhost:0")
-	if err != nil {
-		log.Fatalf("failed to listen: %v", err)
-	}
-
-	srv, err := signalServer.NewServer(context.Background(), otel.Meter(""))
-	require.NoError(t, err)
-	proto.RegisterSignalExchangeServer(s, srv)
-
-	go func() {
-		if err = s.Serve(lis); err != nil {
-			log.Fatalf("failed to serve: %v", err)
-		}
-	}()
-
-	return s, lis.Addr().String(), nil
-}
-
-func startManagement(t *testing.T, dataDir, testFile string) (*grpc.Server, string, error) {
-	t.Helper()
-
-	config := &config.Config{
-		Stuns:      []*config.Host{},
-		TURNConfig: &config.TURNConfig{},
-		Relay: &config.Relay{
-			Addresses:      []string{"127.0.0.1:1234"},
-			CredentialsTTL: util.Duration{Duration: time.Hour},
-			Secret:         "222222222222222222",
-		},
-		Signal: &config.Host{
-			Proto: "http",
-			URI:   "localhost:10000",
-		},
-		Datadir:    dataDir,
-		HttpConfig: nil,
-	}
-
-	lis, err := net.Listen("tcp", "localhost:0")
-	if err != nil {
-		return nil, "", err
-	}
-	s := grpc.NewServer(grpc.KeepaliveEnforcementPolicy(kaep), grpc.KeepaliveParams(kasp))
-
-	store, cleanUp, err := store.NewTestStoreFromSQL(context.Background(), testFile, config.Datadir)
-	if err != nil {
-		return nil, "", err
-	}
-	t.Cleanup(cleanUp)
-
-	eventStore := &activity.InMemoryEventStore{}
-	if err != nil {
-		return nil, "", err
-	}
-
-	permissionsManager := permissions.NewManager(store)
-	peersManager := peers.NewManager(store, permissionsManager)
-	jobManager := job.NewJobManager(nil, store, peersManager)
-
-	cacheStore, err := nbcache.NewStore(context.Background(), 100*time.Millisecond, 300*time.Millisecond, 100)
-	if err != nil {
-		return nil, "", err
-	}
-
-	ia, _ := validator.NewIntegratedValidator(context.Background(), peersManager, nil, eventStore, cacheStore)
-
-	metrics, err := telemetry.NewDefaultAppMetrics(context.Background())
-	require.NoError(t, err)
-
-	ctrl := gomock.NewController(t)
-	t.Cleanup(ctrl.Finish)
-	settingsMockManager := settings.NewMockManager(ctrl)
-	settingsMockManager.EXPECT().
-		GetSettings(gomock.Any(), gomock.Any(), gomock.Any()).
-		Return(&types.Settings{}, nil).
-		AnyTimes()
-	settingsMockManager.EXPECT().
-		GetExtraSettings(gomock.Any(), gomock.Any()).
-		Return(&types.ExtraSettings{}, nil).
-		AnyTimes()
-
-	groupsManager := groups.NewManagerMock()
-
-	updateManager := update_channel.NewPeersUpdateManager(metrics)
-	requestBuffer := server.NewAccountRequestBuffer(context.Background(), store)
-	networkMapController := controller.NewController(context.Background(), store, metrics, updateManager, requestBuffer, server.MockIntegratedValidator{}, settingsMockManager, "netbird.selfhosted", port_forwarding.NewControllerMock(), manager.NewEphemeralManager(store, peersManager), config)
-	accountManager, err := server.BuildManager(context.Background(), config, store, networkMapController, jobManager, nil, "", eventStore, nil, false, ia, metrics, port_forwarding.NewControllerMock(), settingsMockManager, permissionsManager, false, cacheStore)
-	if err != nil {
-		return nil, "", err
-	}
-
-	secretsManager, err := nbgrpc.NewTimeBasedAuthSecretsManager(updateManager, config.TURNConfig, config.Relay, settingsMockManager, groupsManager)
-	if err != nil {
-		return nil, "", err
-	}
-	mgmtServer, err := nbgrpc.NewServer(config, accountManager, settingsMockManager, jobManager, secretsManager, nil, nil, &server.MockIntegratedValidator{}, networkMapController, nil, nil)
-	if err != nil {
-		return nil, "", err
-	}
-	mgmtProto.RegisterManagementServiceServer(s, mgmtServer)
-	go func() {
-		if err = s.Serve(lis); err != nil {
-			log.Fatalf("failed to serve: %v", err)
-		}
-	}()
-
-	return s, lis.Addr().String(), nil
-}
-
-// getConnectedPeers returns a connection Status or nil if peer connection wasn't found
-func getConnectedPeers(e *Engine) int {
-	e.syncMsgMux.Lock()
-	defer e.syncMsgMux.Unlock()
-	i := 0
-	for _, id := range e.peerStore.PeersPubKey() {
-		conn, _ := e.peerStore.PeerConn(id)
-		if conn.IsConnected() {
-			i++
-		}
-	}
-	return i
-}
-
-func getPeers(e *Engine) int {
-	e.syncMsgMux.Lock()
-	defer e.syncMsgMux.Unlock()
-
-	return len(e.peerStore.PeersPubKey())
-}
-
 func mustEncodePrefix(t *testing.T, p netip.Prefix) []byte {
 	t.Helper()
 	b, err := netiputil.EncodePrefix(p)

client/internal/lazyconn/activity/listener_bind.go

@@ -119,10 +119,6 @@ func (d *BindListener) ReadPackets() {
 	}
 
 	d.peerCfg.Log.Debugf("removing lazy endpoint for peer %s", d.peerCfg.PublicKey)
-	if err := d.wgIface.RemovePeer(d.peerCfg.PublicKey); err != nil {
-		d.peerCfg.Log.Errorf("failed to remove endpoint: %s", err)
-	}
-
 	_ = d.lazyConn.Close()
 	d.bind.RemoveEndpoint(d.fakeIP)
 	d.done.Done()

client/internal/mapsync.go

@@ -0,0 +1,190 @@
+package internal
+
+import (
+	"context"
+	"sync"
+	"time"
+
+	log "github.com/sirupsen/logrus"
+
+	mgmProto "github.com/netbirdio/netbird/shared/management/proto"
+)
+
+// mapStateManager is the single read/write point between the management stream
+// (writes) and the convergence loop (reads/applies).
+//
+// The stream calls SetTarget with the latest full SyncResponse — the complete
+// desired state. A single background goroutine (run) applies it to the engine in
+// bounded passes via apply() until converged, releasing syncMsgMux between passes
+// so other subsystems interleave. If a newer update arrives mid-flight, the loop
+// coalesces: it keeps converging toward the latest target and the intermediate one
+// is SKIPPED — never applied on its own (logged, no onConverged).
+//
+// Convergence is a single comparison: appliedGen == targetGen. targetGen
+// increments on every SetTarget (an internal generation counter, so it also covers
+// config-only updates that carry no network-map serial).
+//
+// onConverged fires once for each — and only each — map that is actually processed
+// (i.e. converged as the target). Skipped/superseded maps and dropped-on-error maps
+// do NOT fire it. So "sync finished in X" / RecordSyncDuration always corresponds
+// to a real, completed alignment.
+type mapStateManager struct {
+	// apply performs one bounded apply pass and reports whether more passes are needed.
+	// firstPass is true on the first pass of a given target, so the caller can run
+	// wholesale (firewall/routes/DNS/forward-rules) once per target and skip it on the
+	// re-runs that only drain the bounded peer batches. The manager owns this signal
+	// because it owns the convergence boundary; the engine need not track serials for it.
+	apply func(update *mgmProto.SyncResponse, firstPass bool) (bool, error)
+	// onConverged is called once per processed map, with the elapsed time since that
+	// map was received (for the sync-duration metric / "sync finished" log).
+	onConverged func(time.Duration)
+	// persist snapshots an update to disk for restore-on-restart. Called once per
+	// update received from management (in SetTarget), including ones later coalesced
+	// or skipped from apply, so the on-disk state mirrors what management last sent.
+	// The impl skips config-only updates (nil NetworkMap). May be nil.
+	persist func(*mgmProto.SyncResponse)
+
+	mu          sync.Mutex
+	target      *mgmProto.SyncResponse
+	targetGen   uint64
+	appliedGen  uint64
+	targetSetAt time.Time
+
+	wake chan struct{}
+}
+
+func newMapStateManager(apply func(update *mgmProto.SyncResponse, firstPass bool) (bool, error), persist func(*mgmProto.SyncResponse), onConverged func(time.Duration)) *mapStateManager {
+	return &mapStateManager{
+		apply:       apply,
+		persist:     persist,
+		onConverged: onConverged,
+		wake:        make(chan struct{}, 1),
+	}
+}
+
+// SetTarget records the latest update as the desired state and wakes the loop.
+// It returns immediately; convergence happens in the background. Serial-based
+// staleness of the network map is still enforced inside apply (updateNetworkMap).
+func (m *mapStateManager) SetTarget(update *mgmProto.SyncResponse) error {
+	m.mu.Lock()
+	// A target that has not settled yet (targetGen > appliedGen) is being superseded
+	// before it converged: we coalesce to the latest map and never apply this one on
+	// its own. It is SKIPPED — logged here, and it will not fire onConverged.
+	if m.target != nil && m.targetGen > m.appliedGen {
+		log.Debugf("sync map (gen %d) superseded before convergence, skipping", m.targetGen)
+	}
+	m.target = m.mergeTarget(m.target, update)
+	// Bump an internal generation counter, NOT the map serial: config-only updates
+	// (relay token rotation, STUN/TURN) arrive with NetworkMap == nil and carry no
+	// serial, yet must still be applied. Every SetTarget is therefore a distinct
+	// target regardless of payload. Map-serial staleness is enforced separately
+	// inside apply (updateNetworkMap).
+	m.targetGen++
+	m.targetSetAt = time.Now()
+	m.mu.Unlock()
+
+	select {
+	case m.wake <- struct{}{}:
+	default:
+	}
+
+	// Persist every update received from management — once per update (not per apply
+	// pass), and including ones that get coalesced/skipped from apply, so the on-disk
+	// state always reflects the latest map management sent. Done after waking the loop
+	// so convergence can start in parallel with the disk write. The persist impl skips
+	// config-only updates (nil NetworkMap).
+	if m.persist != nil {
+		m.persist(update)
+	}
+	return nil
+}
+
+// mergeTarget combines the currently pending target with a freshly received update
+// and returns the new desired state. It is called under m.mu from SetTarget and is
+// the single seam where the replace-vs-squash decision lives.
+//
+// Today management always sends a FULL map (the complete desired state), so the
+// update simply replaces whatever was pending — prev is ignored. When management
+// starts sending incremental/delta updates, squash `update` onto `prev` here; the
+// rest of the manager (generation tracking, convergence, signaling) is unaffected
+// because it already treats target as "the complete desired state, whatever it is".
+func (m *mapStateManager) mergeTarget(prev, update *mgmProto.SyncResponse) *mgmProto.SyncResponse {
+	return update
+}
+
+// run drives convergence until ctx is done. It is meant to run in its own goroutine.
+func (m *mapStateManager) run(ctx context.Context) {
+	// passGen is the generation of the most recent apply() call (0 = none). A pass is
+	// the first for its target when its generation differs from the previous one —
+	// true on a fresh target and on a coalesced switch to a newer target mid-flight.
+	var passGen uint64
+	for {
+		m.mu.Lock()
+		target, tg, ag := m.target, m.targetGen, m.appliedGen
+		m.mu.Unlock()
+
+		// Fully converged (or nothing yet): block until a new target arrives.
+		if target == nil || ag == tg {
+			select {
+			case <-ctx.Done():
+				return
+			case <-m.wake:
+				continue
+			}
+		}
+
+		firstPass := tg != passGen
+		passGen = tg
+		more, err := m.apply(target, firstPass)
+		if err != nil {
+			if ctx.Err() != nil {
+				return
+			}
+			// Log and DROP this target — do not retry it. A deterministic failure
+			// (e.g. a malformed peer in the map) would otherwise spin every pass
+			// making no progress. Management is the source of truth and re-delivers
+			// the full map on the next sync, so dropping is safe; peers already
+			// applied this convergence stay (idempotent diffs) and the remainder is
+			// reconciled by the next target. Mirrors the legacy handleSync path,
+			// where the apply error was logged by the gRPC client and the update
+			// dropped. No onConverged: this target did not converge.
+			log.Errorf("apply sync pass, dropping update: %v", err)
+			m.settle(tg, false)
+			continue
+		}
+
+		if more {
+			// keep converging the current target; syncMsgMux was released by apply
+			// between passes so other subsystems interleave.
+			continue
+		}
+
+		// This pass converged. Mark applied and signal this one map.
+		m.settle(tg, true)
+		// if a newer target arrived mid-pass, settle is a no-op (targetGen != tg) and
+		// ag<tg next iteration -> apply it; this generation was skipped (logged in
+		// SetTarget) and is not signaled.
+	}
+}
+
+// settle marks generation tg as processed so the loop goes idle instead of
+// re-applying the same target. It is a no-op when a newer target arrived during the
+// pass (targetGen != tg), leaving appliedGen behind so that target re-applies — the
+// just-finished generation was already counted as skipped.
+//
+// When signal is true (the pass converged) it fires onConverged once for this map;
+// when false (the target was dropped on error) it does not — the map did not converge.
+func (m *mapStateManager) settle(tg uint64, signal bool) {
+	m.mu.Lock()
+	if m.targetGen != tg {
+		m.mu.Unlock()
+		return
+	}
+	m.appliedGen = tg
+	setAt := m.targetSetAt
+	m.mu.Unlock()
+
+	if signal && m.onConverged != nil {
+		m.onConverged(time.Since(setAt))
+	}
+}

client/internal/mapsync_test.go

@@ -0,0 +1,242 @@
+package internal
+
+import (
+	"context"
+	"errors"
+	"sync/atomic"
+	"testing"
+	"time"
+
+	"github.com/stretchr/testify/require"
+
+	mgmProto "github.com/netbirdio/netbird/shared/management/proto"
+)
+
+// converges over the bounded passes (apply returns more until the 3rd pass),
+// fires onConverged exactly once, then blocks (no further apply) until a new target.
+func TestMapStateManager_ConvergesThenStops(t *testing.T) {
+	var passes int32
+	var firstPasses int32
+	converged := make(chan struct{}, 1)
+
+	apply := func(_ *mgmProto.SyncResponse, firstPass bool) (bool, error) {
+		n := atomic.AddInt32(&passes, 1)
+		if firstPass {
+			atomic.AddInt32(&firstPasses, 1)
+		}
+		return n < 3, nil // more on pass 1 and 2, converge on pass 3
+	}
+	m := newMapStateManager(apply, nil, func(time.Duration) { converged <- struct{}{} })
+
+	ctx, cancel := context.WithCancel(context.Background())
+	defer cancel()
+	go m.run(ctx)
+
+	require.NoError(t, m.SetTarget(&mgmProto.SyncResponse{}))
+
+	select {
+	case <-converged:
+	case <-time.After(2 * time.Second):
+		t.Fatal("manager did not converge")
+	}
+	require.EqualValues(t, 3, atomic.LoadInt32(&passes))
+	require.EqualValues(t, 1, atomic.LoadInt32(&firstPasses), "firstPass true only on pass 1, false on re-runs of the same target")
+
+	// once converged the loop blocks: no further apply calls
+	time.Sleep(100 * time.Millisecond)
+	require.EqualValues(t, 3, atomic.LoadInt32(&passes), "apply must not run after convergence")
+}
+
+// persist runs once per received update (not per apply pass), regardless of how many
+// bounded passes that target takes to converge.
+func TestMapStateManager_PersistsOncePerUpdate(t *testing.T) {
+	var passes, persists int32
+	converged := make(chan struct{}, 1)
+	apply := func(_ *mgmProto.SyncResponse, _ bool) (bool, error) {
+		n := atomic.AddInt32(&passes, 1)
+		return n < 3, nil // 3 passes for one target
+	}
+	persist := func(*mgmProto.SyncResponse) { atomic.AddInt32(&persists, 1) }
+	m := newMapStateManager(apply, persist, func(time.Duration) { converged <- struct{}{} })
+
+	ctx, cancel := context.WithCancel(context.Background())
+	defer cancel()
+	go m.run(ctx)
+
+	require.NoError(t, m.SetTarget(&mgmProto.SyncResponse{}))
+	select {
+	case <-converged:
+	case <-time.After(2 * time.Second):
+		t.Fatal("did not converge")
+	}
+	require.EqualValues(t, 3, atomic.LoadInt32(&passes))
+	require.EqualValues(t, 1, atomic.LoadInt32(&persists), "persist once per update, not per pass")
+}
+
+// every update received from management is persisted — even one that is coalesced /
+// skipped from apply before it ever converges.
+func TestMapStateManager_PersistsEveryUpdateIncludingSkipped(t *testing.T) {
+	release := make(chan struct{})
+	var persists int32
+	apply := func(_ *mgmProto.SyncResponse, _ bool) (bool, error) {
+		<-release // hold the first apply so the second update coalesces/skips
+		return false, nil
+	}
+	persist := func(*mgmProto.SyncResponse) { atomic.AddInt32(&persists, 1) }
+	m := newMapStateManager(apply, persist, nil)
+
+	ctx, cancel := context.WithCancel(context.Background())
+	defer cancel()
+	go m.run(ctx)
+
+	require.NoError(t, m.SetTarget(&mgmProto.SyncResponse{})) // map1 -> apply blocks
+	require.NoError(t, m.SetTarget(&mgmProto.SyncResponse{})) // map2 supersedes map1 (skipped from apply)
+	close(release)
+
+	// both updates persisted even though map1 is skipped from apply
+	require.Eventually(t, func() bool { return atomic.LoadInt32(&persists) == 2 }, 2*time.Second, 10*time.Millisecond)
+}
+
+// each map that is actually processed (converged before the next arrives) fires
+// onConverged exactly once — mirroring the legacy per-message handleSync timing.
+func TestMapStateManager_SignalsEachProcessedMap(t *testing.T) {
+	converged := make(chan struct{}, 8)
+	apply := func(_ *mgmProto.SyncResponse, _ bool) (bool, error) {
+		return false, nil // converge in one pass
+	}
+	m := newMapStateManager(apply, nil, func(time.Duration) { converged <- struct{}{} })
+
+	ctx, cancel := context.WithCancel(context.Background())
+	defer cancel()
+	go m.run(ctx)
+
+	const maps = 3
+	for i := 0; i < maps; i++ {
+		require.NoError(t, m.SetTarget(&mgmProto.SyncResponse{}))
+		select { // wait for this map to converge before sending the next (no coalescing)
+		case <-converged:
+		case <-time.After(2 * time.Second):
+			t.Fatalf("map %d not signaled", i)
+		}
+	}
+
+	// no extra signals once the stream goes quiet
+	select {
+	case <-converged:
+		t.Fatal("unexpected extra onConverged")
+	case <-time.After(100 * time.Millisecond):
+	}
+}
+
+// a map superseded before it converges is skipped: only the latest (processed) map
+// fires onConverged, not the skipped one.
+func TestMapStateManager_SkippedMapNotSignaled(t *testing.T) {
+	release := make(chan struct{})
+	var applies, converged atomic.Int32
+	apply := func(_ *mgmProto.SyncResponse, _ bool) (bool, error) {
+		applies.Add(1)
+		<-release // hold the first apply in-flight so we can queue a newer target
+		return false, nil
+	}
+	m := newMapStateManager(apply, nil, func(time.Duration) { converged.Add(1) })
+
+	ctx, cancel := context.WithCancel(context.Background())
+	defer cancel()
+	go m.run(ctx)
+
+	// map1 is picked up; its apply blocks on release
+	require.NoError(t, m.SetTarget(&mgmProto.SyncResponse{}))
+	require.Eventually(t, func() bool { return applies.Load() >= 1 }, 2*time.Second, 5*time.Millisecond)
+
+	// map2 supersedes map1 before it settled -> map1 is skipped
+	require.NoError(t, m.SetTarget(&mgmProto.SyncResponse{}))
+	close(release) // let both applies proceed
+
+	// only the processed (latest) map signals; the skipped one does not
+	require.Eventually(t, func() bool { return converged.Load() == 1 }, 2*time.Second, 10*time.Millisecond)
+	time.Sleep(150 * time.Millisecond)
+	require.EqualValues(t, 1, converged.Load(), "skipped map must not fire onConverged")
+	require.EqualValues(t, 2, applies.Load(), "both targets entered apply (map1 once, map2 once)")
+}
+
+// an apply error drops the target: no retry of the same target, no onConverged,
+// the loop goes idle — and a fresh target is still applied afterwards.
+func TestMapStateManager_DropsTargetOnError(t *testing.T) {
+	applied := make(chan struct{}, 8)
+	var failNext atomic.Bool
+	failNext.Store(true)
+
+	apply := func(_ *mgmProto.SyncResponse, _ bool) (bool, error) {
+		applied <- struct{}{}
+		if failNext.Load() {
+			return false, errors.New("boom")
+		}
+		return false, nil // converge in one pass
+	}
+	var converged atomic.Int32
+	m := newMapStateManager(apply, nil, func(time.Duration) { converged.Add(1) })
+
+	ctx, cancel := context.WithCancel(context.Background())
+	defer cancel()
+	go m.run(ctx)
+
+	// first target errors -> applied once, then dropped (no retry, no onConverged)
+	require.NoError(t, m.SetTarget(&mgmProto.SyncResponse{}))
+	select {
+	case <-applied:
+	case <-time.After(2 * time.Second):
+		t.Fatal("errored target not applied")
+	}
+	select {
+	case <-applied:
+		t.Fatal("errored target must not be retried")
+	case <-time.After(150 * time.Millisecond):
+	}
+	require.EqualValues(t, 0, converged.Load(), "onConverged must not fire on error")
+
+	// a new target is still processed normally and converges
+	failNext.Store(false)
+	require.NoError(t, m.SetTarget(&mgmProto.SyncResponse{}))
+	select {
+	case <-applied:
+	case <-time.After(2 * time.Second):
+		t.Fatal("new target after error not applied")
+	}
+	require.Eventually(t, func() bool { return converged.Load() == 1 }, 2*time.Second, 10*time.Millisecond)
+}
+
+// a new target after convergence triggers a fresh apply; an idle (converged)
+// manager does not apply on its own.
+func TestMapStateManager_ReappliesOnNewTarget(t *testing.T) {
+	applied := make(chan struct{}, 8)
+	apply := func(_ *mgmProto.SyncResponse, _ bool) (bool, error) {
+		applied <- struct{}{}
+		return false, nil // converge in one pass
+	}
+	m := newMapStateManager(apply, nil, nil)
+
+	ctx, cancel := context.WithCancel(context.Background())
+	defer cancel()
+	go m.run(ctx)
+
+	require.NoError(t, m.SetTarget(&mgmProto.SyncResponse{}))
+	select {
+	case <-applied:
+	case <-time.After(2 * time.Second):
+		t.Fatal("first target not applied")
+	}
+
+	// converged → must stay idle (no spurious apply)
+	select {
+	case <-applied:
+		t.Fatal("unexpected apply while idle/converged")
+	case <-time.After(150 * time.Millisecond):
+	}
+
+	require.NoError(t, m.SetTarget(&mgmProto.SyncResponse{}))
+	select {
+	case <-applied:
+	case <-time.After(2 * time.Second):
+		t.Fatal("new target not applied")
+	}
+}

client/internal/peer/handshaker.go

@@ -195,14 +195,14 @@ func (h *Handshaker) sendOffer() error {
 	}
 
 	offer := h.buildOfferAnswer()
-	h.log.Infof("sending offer with serial: %s", offer.SessionIDString())
+	h.log.Debugf("sending offer with serial: %s", offer.SessionIDString())
 
 	return h.signaler.SignalOffer(offer, h.config.Key)
 }
 
 func (h *Handshaker) sendAnswer() error {
 	answer := h.buildOfferAnswer()
-	h.log.Infof("sending answer with serial: %s", answer.SessionIDString())
+	h.log.Debugf("sending answer with serial: %s", answer.SessionIDString())
 
 	return h.signaler.SignalAnswer(answer, h.config.Key)
 }

client/internal/peer/status.go

@@ -192,6 +192,7 @@ func (s *StatusChangeSubscription) Events() chan map[string]RouterState {
 // Pure read methods take RLock; anything that mutates state takes Lock.
 type Status struct {
 	mux                   sync.RWMutex
+	muxRelays             sync.RWMutex
 	peers                 map[string]State
 	ipToKey               map[string]string
 	changeNotify          map[string]map[string]*StatusChangeSubscription // map[peerID]map[subscriptionID]*StatusChangeSubscription
@@ -244,8 +245,8 @@ func NewRecorder(mgmAddress string) *Status {
 }
 
 func (d *Status) SetRelayMgr(manager *relayClient.Manager) {
-	d.mux.Lock()
+	d.muxRelays.Lock()
-	defer d.mux.Unlock()
+	defer d.muxRelays.Unlock()
 	d.relayMgr = manager
 }
 
@@ -906,8 +907,8 @@ func (d *Status) MarkSignalConnected() {
 }
 
 func (d *Status) UpdateRelayStates(relayResults []relay.ProbeResult) {
-	d.mux.Lock()
+	d.muxRelays.Lock()
-	defer d.mux.Unlock()
+	defer d.muxRelays.Unlock()
 	d.relayStates = relayResults
 }
 
@@ -1018,24 +1019,26 @@ func (d *Status) GetSignalState() SignalState {
 
 // GetRelayStates returns the stun/turn/permanent relay states
 func (d *Status) GetRelayStates() []relay.ProbeResult {
-	d.mux.RLock()
+	d.muxRelays.RLock()
-	defer d.mux.RUnlock()
 	if d.relayMgr == nil {
-		return d.relayStates
+		defer d.muxRelays.RUnlock()
+		return slices.Clone(d.relayStates)
 	}
 
+	relayMgr := d.relayMgr
 	// extend the list of stun, turn servers with the relay server connections
 	relayStates := slices.Clone(d.relayStates)
+	d.muxRelays.RUnlock()
 
-	states := d.relayMgr.RelayStates()
+	states := relayMgr.RelayStates()
 	if len(states) == 0 {
 		// no relay connection tracked yet; surface configured servers as
 		// unavailable with the real reconnect error when known
 		err := relayClient.ErrRelayClientNotConnected
-		if connErr := d.relayMgr.RelayConnectError(); connErr != nil {
+		if connErr := relayMgr.RelayConnectError(); connErr != nil {
 			err = connErr
 		}
-		for _, r := range d.relayMgr.ServerURLs() {
+		for _, r := range relayMgr.ServerURLs() {
 			relayStates = append(relayStates, relay.ProbeResult{
 				URI: r,
 				Err: err,

client/internal/profilemanager/config.go

@@ -433,7 +433,7 @@ func (config *Config) apply(input ConfigInput) (updated bool, err error) {
 		updated = true
 	}
 
-	if input.ServerSSHAllowed != nil && *input.ServerSSHAllowed != *config.ServerSSHAllowed {
+	if input.ServerSSHAllowed != nil && (config.ServerSSHAllowed == nil || *input.ServerSSHAllowed != *config.ServerSSHAllowed) {
 		if *input.ServerSSHAllowed {
 			log.Infof("enabling SSH server")
 		} else {

client/internal/profilemanager/config_test.go

@@ -242,6 +242,35 @@ func TestWireguardPortDefaultVsExplicit(t *testing.T) {
 	}
 }
 
+func TestUpdateConfigServerSSHAllowedNotSet(t *testing.T) {
+	// Configs written before ServerSSHAllowed was introduced lack the field and
+	// unmarshal to nil. Supplying the SSH server flag on top of such a config must
+	// apply the value instead of panicking on a nil pointer dereference.
+	tests := []struct {
+		name  string
+		input *bool
+		want  bool
+	}{
+		{"enable", util.True(), true},
+		{"disable", util.False(), false},
+	}
+
+	for _, tt := range tests {
+		t.Run(tt.name, func(t *testing.T) {
+			configPath := filepath.Join(t.TempDir(), "config.json")
+			require.NoError(t, os.WriteFile(configPath, []byte("{}"), 0600))
+
+			config, err := UpdateConfig(ConfigInput{
+				ConfigPath:       configPath,
+				ServerSSHAllowed: tt.input,
+			})
+			require.NoError(t, err)
+			require.NotNil(t, config.ServerSSHAllowed, "ServerSSHAllowed should be set from input")
+			assert.Equal(t, tt.want, *config.ServerSSHAllowed)
+		})
+	}
+}
+
 func TestUpdateOldManagementURL(t *testing.T) {
 	origProber := newMgmProber
 	newMgmProber = func(_ context.Context, _ string, _ wgtypes.Key, _ bool) (mgmProber, error) {

client/internal/routemanager/manager_test.go

@@ -1,3 +1,5 @@
+//go:build privileged
+
 package routemanager
 
 import (

client/internal/routemanager/systemops/rt_tables_linux_test.go

@@ -0,0 +1,69 @@
+//go:build linux && !android
+
+package systemops
+
+import (
+	"fmt"
+	"os"
+	"strings"
+	"testing"
+
+	"github.com/stretchr/testify/assert"
+	"github.com/stretchr/testify/require"
+)
+
+func TestEntryExists(t *testing.T) {
+	tempDir := t.TempDir()
+	tempFilePath := fmt.Sprintf("%s/rt_tables", tempDir)
+
+	content := []string{
+		"1000 reserved",
+		fmt.Sprintf("%d %s", NetbirdVPNTableID, NetbirdVPNTableName),
+		"9999 other_table",
+	}
+	require.NoError(t, os.WriteFile(tempFilePath, []byte(strings.Join(content, "\n")), 0644))
+
+	file, err := os.Open(tempFilePath)
+	require.NoError(t, err)
+	defer func() {
+		assert.NoError(t, file.Close())
+	}()
+
+	tests := []struct {
+		name        string
+		id          int
+		shouldExist bool
+		err         error
+	}{
+		{
+			name:        "ExistsWithNetbirdPrefix",
+			id:          7120,
+			shouldExist: true,
+			err:         nil,
+		},
+		{
+			name:        "ExistsWithDifferentName",
+			id:          1000,
+			shouldExist: true,
+			err:         ErrTableIDExists,
+		},
+		{
+			name:        "DoesNotExist",
+			id:          1234,
+			shouldExist: false,
+			err:         nil,
+		},
+	}
+
+	for _, tc := range tests {
+		t.Run(tc.name, func(t *testing.T) {
+			exists, err := entryExists(file, tc.id)
+			if tc.err != nil {
+				assert.ErrorIs(t, err, tc.err)
+			} else {
+				assert.NoError(t, err)
+			}
+			assert.Equal(t, tc.shouldExist, exists)
+		})
+	}
+}

client/internal/routemanager/systemops/systemops_bsd_privileged_test.go

@@ -0,0 +1,191 @@
+//go:build (darwin || dragonfly || freebsd || netbsd || openbsd) && privileged
+
+package systemops
+
+import (
+	"fmt"
+	"net"
+	"net/netip"
+	"os/exec"
+	"regexp"
+	"runtime"
+	"strings"
+	"sync"
+	"testing"
+
+	"github.com/stretchr/testify/assert"
+	"github.com/stretchr/testify/require"
+)
+
+func init() {
+	testCases = append(testCases, []testCase{
+		{
+			name:              "To more specific route without custom dialer via vpn",
+			expectedInterface: expectedVPNint,
+			dialer:            &net.Dialer{},
+			expectedPacket:    createPacketExpectation("100.64.0.1", 12345, "10.10.0.2", 53),
+		},
+	}...)
+}
+
+func TestConcurrentRoutes(t *testing.T) {
+	baseIP := netip.MustParseAddr("192.0.2.0")
+
+	var intf *net.Interface
+	var nexthop Nexthop
+
+	_, intf = setupDummyInterface(t)
+	nexthop = Nexthop{netip.Addr{}, intf}
+
+	r := New(nil, nil)
+
+	var wg sync.WaitGroup
+	for i := 0; i < 1024; i++ {
+		wg.Add(1)
+		go func(ip netip.Addr) {
+			defer wg.Done()
+			prefix := netip.PrefixFrom(ip, 32)
+			if err := r.addToRouteTable(prefix, nexthop); err != nil {
+				t.Errorf("Failed to add route for %s: %v", prefix, err)
+			}
+		}(baseIP)
+		baseIP = baseIP.Next()
+	}
+
+	wg.Wait()
+
+	baseIP = netip.MustParseAddr("192.0.2.0")
+
+	for i := 0; i < 1024; i++ {
+		wg.Add(1)
+		go func(ip netip.Addr) {
+			defer wg.Done()
+			prefix := netip.PrefixFrom(ip, 32)
+			if err := r.removeFromRouteTable(prefix, nexthop); err != nil {
+				t.Errorf("Failed to remove route for %s: %v", prefix, err)
+			}
+		}(baseIP)
+		baseIP = baseIP.Next()
+	}
+
+	wg.Wait()
+}
+
+func createAndSetupDummyInterface(t *testing.T, intf string, ipAddressCIDR string) string {
+	t.Helper()
+
+	if runtime.GOOS == "darwin" {
+		err := exec.Command("ifconfig", intf, "alias", ipAddressCIDR).Run()
+		require.NoError(t, err, "Failed to create loopback alias")
+
+		t.Cleanup(func() {
+			err := exec.Command("ifconfig", intf, ipAddressCIDR, "-alias").Run()
+			assert.NoError(t, err, "Failed to remove loopback alias")
+		})
+
+		return intf
+	}
+
+	prefix, err := netip.ParsePrefix(ipAddressCIDR)
+	require.NoError(t, err, "Failed to parse prefix")
+
+	netIntf, err := net.InterfaceByName(intf)
+	require.NoError(t, err, "Failed to get interface by name")
+
+	nexthop := Nexthop{netip.Addr{}, netIntf}
+
+	r := New(nil, nil)
+	err = r.addToRouteTable(prefix, nexthop)
+	require.NoError(t, err, "Failed to add route to table")
+
+	t.Cleanup(func() {
+		err := r.removeFromRouteTable(prefix, nexthop)
+		assert.NoError(t, err, "Failed to remove route from table")
+	})
+
+	return intf
+}
+
+func addDummyRoute(t *testing.T, dstCIDR string, gw netip.Addr, _ string) {
+	t.Helper()
+
+	var originalNexthop net.IP
+	if dstCIDR == "0.0.0.0/0" {
+		var err error
+		originalNexthop, err = fetchOriginalGateway()
+		if err != nil {
+			t.Logf("Failed to fetch original gateway: %v", err)
+		}
+
+		if output, err := exec.Command("route", "delete", "-net", dstCIDR).CombinedOutput(); err != nil {
+			t.Logf("Failed to delete route: %v, output: %s", err, output)
+		}
+	}
+
+	t.Cleanup(func() {
+		if originalNexthop != nil {
+			err := exec.Command("route", "add", "-net", dstCIDR, originalNexthop.String()).Run()
+			assert.NoError(t, err, "Failed to restore original route")
+		}
+	})
+
+	err := exec.Command("route", "add", "-net", dstCIDR, gw.String()).Run()
+	require.NoError(t, err, "Failed to add route")
+
+	t.Cleanup(func() {
+		err := exec.Command("route", "delete", "-net", dstCIDR).Run()
+		assert.NoError(t, err, "Failed to remove route")
+	})
+}
+
+func fetchOriginalGateway() (net.IP, error) {
+	output, err := exec.Command("route", "-n", "get", "default").CombinedOutput()
+	if err != nil {
+		return nil, err
+	}
+
+	matches := regexp.MustCompile(`gateway: (\S+)`).FindStringSubmatch(string(output))
+	if len(matches) == 0 {
+		return nil, fmt.Errorf("gateway not found")
+	}
+
+	return net.ParseIP(matches[1]), nil
+}
+
+// setupDummyInterface creates a dummy tun interface for FreeBSD route testing
+func setupDummyInterface(t *testing.T) (netip.Addr, *net.Interface) {
+	t.Helper()
+
+	if runtime.GOOS == "darwin" {
+		return netip.AddrFrom4([4]byte{192, 168, 1, 2}), &net.Interface{Name: "lo0"}
+	}
+
+	output, err := exec.Command("ifconfig", "tun", "create").CombinedOutput()
+	require.NoError(t, err, "Failed to create tun interface: %s", string(output))
+
+	tunName := strings.TrimSpace(string(output))
+
+	output, err = exec.Command("ifconfig", tunName, "192.168.1.1", "netmask", "255.255.0.0", "192.168.1.2", "up").CombinedOutput()
+	require.NoError(t, err, "Failed to configure tun interface: %s", string(output))
+
+	intf, err := net.InterfaceByName(tunName)
+	require.NoError(t, err, "Failed to get interface by name")
+
+	t.Cleanup(func() {
+		if err := exec.Command("ifconfig", tunName, "destroy").Run(); err != nil {
+			t.Logf("Failed to destroy tun interface %s: %v", tunName, err)
+		}
+	})
+
+	return netip.AddrFrom4([4]byte{192, 168, 1, 2}), intf
+}
+
+func setupDummyInterfacesAndRoutes(t *testing.T) {
+	t.Helper()
+
+	defaultDummy := createAndSetupDummyInterface(t, expectedExternalInt, "192.168.0.1/24")
+	addDummyRoute(t, "0.0.0.0/0", netip.AddrFrom4([4]byte{192, 168, 0, 1}), defaultDummy)
+
+	otherDummy := createAndSetupDummyInterface(t, expectedInternalInt, "192.168.1.1/24")
+	addDummyRoute(t, "10.0.0.0/8", netip.AddrFrom4([4]byte{192, 168, 1, 1}), otherDummy)
+}

client/internal/routemanager/systemops/systemops_bsd_test.go

@@ -3,79 +3,24 @@
 package systemops
 
 import (
-	"fmt"
-	"net"
-	"net/netip"
-	"os/exec"
-	"regexp"
-	"runtime"
-	"strings"
-	"sync"
 	"testing"
 
 	"github.com/stretchr/testify/assert"
-	"github.com/stretchr/testify/require"
 	"golang.org/x/net/route"
 )
 
+// Interface names used by the shared routing test fixtures. Kept untagged (no
+// privileged build tag) so the non-privileged test files in this package compile.
+//
+//nolint:unused // consumed by the privileged-tagged routing tests
 var expectedVPNint = "utun100"
+
+//nolint:unused // consumed by the privileged-tagged routing tests
 var expectedExternalInt = "lo0"
+
+//nolint:unused // consumed by the privileged-tagged routing tests
 var expectedInternalInt = "lo0"
 
-func init() {
-	testCases = append(testCases, []testCase{
-		{
-			name:              "To more specific route without custom dialer via vpn",
-			expectedInterface: expectedVPNint,
-			dialer:            &net.Dialer{},
-			expectedPacket:    createPacketExpectation("100.64.0.1", 12345, "10.10.0.2", 53),
-		},
-	}...)
-}
-
-func TestConcurrentRoutes(t *testing.T) {
-	baseIP := netip.MustParseAddr("192.0.2.0")
-
-	var intf *net.Interface
-	var nexthop Nexthop
-
-	_, intf = setupDummyInterface(t)
-	nexthop = Nexthop{netip.Addr{}, intf}
-
-	r := New(nil, nil)
-
-	var wg sync.WaitGroup
-	for i := 0; i < 1024; i++ {
-		wg.Add(1)
-		go func(ip netip.Addr) {
-			defer wg.Done()
-			prefix := netip.PrefixFrom(ip, 32)
-			if err := r.addToRouteTable(prefix, nexthop); err != nil {
-				t.Errorf("Failed to add route for %s: %v", prefix, err)
-			}
-		}(baseIP)
-		baseIP = baseIP.Next()
-	}
-
-	wg.Wait()
-
-	baseIP = netip.MustParseAddr("192.0.2.0")
-
-	for i := 0; i < 1024; i++ {
-		wg.Add(1)
-		go func(ip netip.Addr) {
-			defer wg.Done()
-			prefix := netip.PrefixFrom(ip, 32)
-			if err := r.removeFromRouteTable(prefix, nexthop); err != nil {
-				t.Errorf("Failed to remove route for %s: %v", prefix, err)
-			}
-		}(baseIP)
-		baseIP = baseIP.Next()
-	}
-
-	wg.Wait()
-}
-
 func TestBits(t *testing.T) {
 	tests := []struct {
 		name    string
@@ -122,122 +67,3 @@ func TestBits(t *testing.T) {
 		})
 	}
 }
-
-func createAndSetupDummyInterface(t *testing.T, intf string, ipAddressCIDR string) string {
-	t.Helper()
-
-	if runtime.GOOS == "darwin" {
-		err := exec.Command("ifconfig", intf, "alias", ipAddressCIDR).Run()
-		require.NoError(t, err, "Failed to create loopback alias")
-
-		t.Cleanup(func() {
-			err := exec.Command("ifconfig", intf, ipAddressCIDR, "-alias").Run()
-			assert.NoError(t, err, "Failed to remove loopback alias")
-		})
-
-		return intf
-	}
-
-	prefix, err := netip.ParsePrefix(ipAddressCIDR)
-	require.NoError(t, err, "Failed to parse prefix")
-
-	netIntf, err := net.InterfaceByName(intf)
-	require.NoError(t, err, "Failed to get interface by name")
-
-	nexthop := Nexthop{netip.Addr{}, netIntf}
-
-	r := New(nil, nil)
-	err = r.addToRouteTable(prefix, nexthop)
-	require.NoError(t, err, "Failed to add route to table")
-
-	t.Cleanup(func() {
-		err := r.removeFromRouteTable(prefix, nexthop)
-		assert.NoError(t, err, "Failed to remove route from table")
-	})
-
-	return intf
-}
-
-func addDummyRoute(t *testing.T, dstCIDR string, gw netip.Addr, _ string) {
-	t.Helper()
-
-	var originalNexthop net.IP
-	if dstCIDR == "0.0.0.0/0" {
-		var err error
-		originalNexthop, err = fetchOriginalGateway()
-		if err != nil {
-			t.Logf("Failed to fetch original gateway: %v", err)
-		}
-
-		if output, err := exec.Command("route", "delete", "-net", dstCIDR).CombinedOutput(); err != nil {
-			t.Logf("Failed to delete route: %v, output: %s", err, output)
-		}
-	}
-
-	t.Cleanup(func() {
-		if originalNexthop != nil {
-			err := exec.Command("route", "add", "-net", dstCIDR, originalNexthop.String()).Run()
-			assert.NoError(t, err, "Failed to restore original route")
-		}
-	})
-
-	err := exec.Command("route", "add", "-net", dstCIDR, gw.String()).Run()
-	require.NoError(t, err, "Failed to add route")
-
-	t.Cleanup(func() {
-		err := exec.Command("route", "delete", "-net", dstCIDR).Run()
-		assert.NoError(t, err, "Failed to remove route")
-	})
-}
-
-func fetchOriginalGateway() (net.IP, error) {
-	output, err := exec.Command("route", "-n", "get", "default").CombinedOutput()
-	if err != nil {
-		return nil, err
-	}
-
-	matches := regexp.MustCompile(`gateway: (\S+)`).FindStringSubmatch(string(output))
-	if len(matches) == 0 {
-		return nil, fmt.Errorf("gateway not found")
-	}
-
-	return net.ParseIP(matches[1]), nil
-}
-
-// setupDummyInterface creates a dummy tun interface for FreeBSD route testing
-func setupDummyInterface(t *testing.T) (netip.Addr, *net.Interface) {
-	t.Helper()
-
-	if runtime.GOOS == "darwin" {
-		return netip.AddrFrom4([4]byte{192, 168, 1, 2}), &net.Interface{Name: "lo0"}
-	}
-
-	output, err := exec.Command("ifconfig", "tun", "create").CombinedOutput()
-	require.NoError(t, err, "Failed to create tun interface: %s", string(output))
-
-	tunName := strings.TrimSpace(string(output))
-
-	output, err = exec.Command("ifconfig", tunName, "192.168.1.1", "netmask", "255.255.0.0", "192.168.1.2", "up").CombinedOutput()
-	require.NoError(t, err, "Failed to configure tun interface: %s", string(output))
-
-	intf, err := net.InterfaceByName(tunName)
-	require.NoError(t, err, "Failed to get interface by name")
-
-	t.Cleanup(func() {
-		if err := exec.Command("ifconfig", tunName, "destroy").Run(); err != nil {
-			t.Logf("Failed to destroy tun interface %s: %v", tunName, err)
-		}
-	})
-
-	return netip.AddrFrom4([4]byte{192, 168, 1, 2}), intf
-}
-
-func setupDummyInterfacesAndRoutes(t *testing.T) {
-	t.Helper()
-
-	defaultDummy := createAndSetupDummyInterface(t, expectedExternalInt, "192.168.0.1/24")
-	addDummyRoute(t, "0.0.0.0/0", netip.AddrFrom4([4]byte{192, 168, 0, 1}), defaultDummy)
-
-	otherDummy := createAndSetupDummyInterface(t, expectedInternalInt, "192.168.1.1/24")
-	addDummyRoute(t, "10.0.0.0/8", netip.AddrFrom4([4]byte{192, 168, 1, 1}), otherDummy)
-}

client/internal/routemanager/systemops/systemops_dialer_test.go

@@ -0,0 +1,17 @@
+//go:build !android && !ios
+
+package systemops
+
+import (
+	"context"
+	"net"
+)
+
+// dialer is shared by the per-platform routing test cases. Kept untagged (no
+// privileged build tag) so the non-privileged test files compile on every platform.
+//
+//nolint:unused // consumed by the privileged-tagged routing tests
+type dialer interface {
+	Dial(network, address string) (net.Conn, error)
+	DialContext(ctx context.Context, network, address string) (net.Conn, error)
+}

client/internal/routemanager/systemops/systemops_generic_test.go

@@ -1,4 +1,4 @@
-//go:build !android && !ios
+//go:build !android && !ios && privileged
 
 package systemops
 
@@ -26,11 +26,6 @@ import (
 	nbnet "github.com/netbirdio/netbird/client/net"
 )
 
-type dialer interface {
-	Dial(network, address string) (net.Conn, error)
-	DialContext(ctx context.Context, network, address string) (net.Conn, error)
-}
-
 func TestAddVPNRoute(t *testing.T) {
 	testCases := []struct {
 		name        string
@@ -515,125 +510,3 @@ func setupTestEnv(t *testing.T) {
 	// unique route in vpn table
 	setupRouteAndCleanup(t, r, netip.MustParsePrefix("172.16.0.0/12"), intf)
 }
-
-func TestIsVpnRoute(t *testing.T) {
-	tests := []struct {
-		name           string
-		addr           string
-		vpnRoutes      []string
-		localRoutes    []string
-		expectedVpn    bool
-		expectedPrefix netip.Prefix
-	}{
-		{
-			name:           "Match in VPN routes",
-			addr:           "192.168.1.1",
-			vpnRoutes:      []string{"192.168.1.0/24"},
-			localRoutes:    []string{"10.0.0.0/8"},
-			expectedVpn:    true,
-			expectedPrefix: netip.MustParsePrefix("192.168.1.0/24"),
-		},
-		{
-			name:           "Match in local routes",
-			addr:           "10.1.1.1",
-			vpnRoutes:      []string{"192.168.1.0/24"},
-			localRoutes:    []string{"10.0.0.0/8"},
-			expectedVpn:    false,
-			expectedPrefix: netip.MustParsePrefix("10.0.0.0/8"),
-		},
-		{
-			name:           "No match",
-			addr:           "172.16.0.1",
-			vpnRoutes:      []string{"192.168.1.0/24"},
-			localRoutes:    []string{"10.0.0.0/8"},
-			expectedVpn:    false,
-			expectedPrefix: netip.Prefix{},
-		},
-		{
-			name:           "Default route ignored",
-			addr:           "192.168.1.1",
-			vpnRoutes:      []string{"0.0.0.0/0", "192.168.1.0/24"},
-			localRoutes:    []string{"10.0.0.0/8"},
-			expectedVpn:    true,
-			expectedPrefix: netip.MustParsePrefix("192.168.1.0/24"),
-		},
-		{
-			name:           "Default route matches but ignored",
-			addr:           "172.16.1.1",
-			vpnRoutes:      []string{"0.0.0.0/0", "192.168.1.0/24"},
-			localRoutes:    []string{"10.0.0.0/8"},
-			expectedVpn:    false,
-			expectedPrefix: netip.Prefix{},
-		},
-		{
-			name:           "Longest prefix match local",
-			addr:           "192.168.1.1",
-			vpnRoutes:      []string{"192.168.0.0/16"},
-			localRoutes:    []string{"192.168.1.0/24"},
-			expectedVpn:    false,
-			expectedPrefix: netip.MustParsePrefix("192.168.1.0/24"),
-		},
-		{
-			name:           "Longest prefix match local multiple",
-			addr:           "192.168.0.1",
-			vpnRoutes:      []string{"192.168.0.0/16", "192.168.0.0/25", "192.168.0.0/27"},
-			localRoutes:    []string{"192.168.0.0/24", "192.168.0.0/26", "192.168.0.0/28"},
-			expectedVpn:    false,
-			expectedPrefix: netip.MustParsePrefix("192.168.0.0/28"),
-		},
-		{
-			name:           "Longest prefix match vpn",
-			addr:           "192.168.1.1",
-			vpnRoutes:      []string{"192.168.1.0/24"},
-			localRoutes:    []string{"192.168.0.0/16"},
-			expectedVpn:    true,
-			expectedPrefix: netip.MustParsePrefix("192.168.1.0/24"),
-		},
-		{
-			name:           "Longest prefix match vpn multiple",
-			addr:           "192.168.0.1",
-			vpnRoutes:      []string{"192.168.0.0/16", "192.168.0.0/25", "192.168.0.0/27"},
-			localRoutes:    []string{"192.168.0.0/24", "192.168.0.0/26"},
-			expectedVpn:    true,
-			expectedPrefix: netip.MustParsePrefix("192.168.0.0/27"),
-		},
-		{
-			name:           "Duplicate prefix in both",
-			addr:           "192.168.1.1",
-			vpnRoutes:      []string{"192.168.1.0/24"},
-			localRoutes:    []string{"192.168.1.0/24"},
-			expectedVpn:    false,
-			expectedPrefix: netip.MustParsePrefix("192.168.1.0/24"),
-		},
-	}
-
-	for _, tt := range tests {
-		t.Run(tt.name, func(t *testing.T) {
-			addr, err := netip.ParseAddr(tt.addr)
-			if err != nil {
-				t.Fatalf("Failed to parse address %s: %v", tt.addr, err)
-			}
-
-			var vpnRoutes, localRoutes []netip.Prefix
-			for _, route := range tt.vpnRoutes {
-				prefix, err := netip.ParsePrefix(route)
-				if err != nil {
-					t.Fatalf("Failed to parse VPN route %s: %v", route, err)
-				}
-				vpnRoutes = append(vpnRoutes, prefix)
-			}
-
-			for _, route := range tt.localRoutes {
-				prefix, err := netip.ParsePrefix(route)
-				if err != nil {
-					t.Fatalf("Failed to parse local route %s: %v", route, err)
-				}
-				localRoutes = append(localRoutes, prefix)
-			}
-
-			isVpn, matchedPrefix := isVpnRoute(addr, vpnRoutes, localRoutes)
-			assert.Equal(t, tt.expectedVpn, isVpn, "isVpnRoute should return expectedVpn value")
-			assert.Equal(t, tt.expectedPrefix, matchedPrefix, "isVpnRoute should return expectedVpn prefix")
-		})
-	}
-}

client/internal/routemanager/systemops/systemops_isvpnroute_test.go

@@ -0,0 +1,132 @@
+//go:build !android && !ios
+
+package systemops
+
+import (
+	"net/netip"
+	"testing"
+
+	"github.com/stretchr/testify/assert"
+)
+
+func TestIsVpnRoute(t *testing.T) {
+	tests := []struct {
+		name           string
+		addr           string
+		vpnRoutes      []string
+		localRoutes    []string
+		expectedVpn    bool
+		expectedPrefix netip.Prefix
+	}{
+		{
+			name:           "Match in VPN routes",
+			addr:           "192.168.1.1",
+			vpnRoutes:      []string{"192.168.1.0/24"},
+			localRoutes:    []string{"10.0.0.0/8"},
+			expectedVpn:    true,
+			expectedPrefix: netip.MustParsePrefix("192.168.1.0/24"),
+		},
+		{
+			name:           "Match in local routes",
+			addr:           "10.1.1.1",
+			vpnRoutes:      []string{"192.168.1.0/24"},
+			localRoutes:    []string{"10.0.0.0/8"},
+			expectedVpn:    false,
+			expectedPrefix: netip.MustParsePrefix("10.0.0.0/8"),
+		},
+		{
+			name:           "No match",
+			addr:           "172.16.0.1",
+			vpnRoutes:      []string{"192.168.1.0/24"},
+			localRoutes:    []string{"10.0.0.0/8"},
+			expectedVpn:    false,
+			expectedPrefix: netip.Prefix{},
+		},
+		{
+			name:           "Default route ignored",
+			addr:           "192.168.1.1",
+			vpnRoutes:      []string{"0.0.0.0/0", "192.168.1.0/24"},
+			localRoutes:    []string{"10.0.0.0/8"},
+			expectedVpn:    true,
+			expectedPrefix: netip.MustParsePrefix("192.168.1.0/24"),
+		},
+		{
+			name:           "Default route matches but ignored",
+			addr:           "172.16.1.1",
+			vpnRoutes:      []string{"0.0.0.0/0", "192.168.1.0/24"},
+			localRoutes:    []string{"10.0.0.0/8"},
+			expectedVpn:    false,
+			expectedPrefix: netip.Prefix{},
+		},
+		{
+			name:           "Longest prefix match local",
+			addr:           "192.168.1.1",
+			vpnRoutes:      []string{"192.168.0.0/16"},
+			localRoutes:    []string{"192.168.1.0/24"},
+			expectedVpn:    false,
+			expectedPrefix: netip.MustParsePrefix("192.168.1.0/24"),
+		},
+		{
+			name:           "Longest prefix match local multiple",
+			addr:           "192.168.0.1",
+			vpnRoutes:      []string{"192.168.0.0/16", "192.168.0.0/25", "192.168.0.0/27"},
+			localRoutes:    []string{"192.168.0.0/24", "192.168.0.0/26", "192.168.0.0/28"},
+			expectedVpn:    false,
+			expectedPrefix: netip.MustParsePrefix("192.168.0.0/28"),
+		},
+		{
+			name:           "Longest prefix match vpn",
+			addr:           "192.168.1.1",
+			vpnRoutes:      []string{"192.168.1.0/24"},
+			localRoutes:    []string{"192.168.0.0/16"},
+			expectedVpn:    true,
+			expectedPrefix: netip.MustParsePrefix("192.168.1.0/24"),
+		},
+		{
+			name:           "Longest prefix match vpn multiple",
+			addr:           "192.168.0.1",
+			vpnRoutes:      []string{"192.168.0.0/16", "192.168.0.0/25", "192.168.0.0/27"},
+			localRoutes:    []string{"192.168.0.0/24", "192.168.0.0/26"},
+			expectedVpn:    true,
+			expectedPrefix: netip.MustParsePrefix("192.168.0.0/27"),
+		},
+		{
+			name:           "Duplicate prefix in both",
+			addr:           "192.168.1.1",
+			vpnRoutes:      []string{"192.168.1.0/24"},
+			localRoutes:    []string{"192.168.1.0/24"},
+			expectedVpn:    false,
+			expectedPrefix: netip.MustParsePrefix("192.168.1.0/24"),
+		},
+	}
+
+	for _, tt := range tests {
+		t.Run(tt.name, func(t *testing.T) {
+			addr, err := netip.ParseAddr(tt.addr)
+			if err != nil {
+				t.Fatalf("Failed to parse address %s: %v", tt.addr, err)
+			}
+
+			var vpnRoutes, localRoutes []netip.Prefix
+			for _, route := range tt.vpnRoutes {
+				prefix, err := netip.ParsePrefix(route)
+				if err != nil {
+					t.Fatalf("Failed to parse VPN route %s: %v", route, err)
+				}
+				vpnRoutes = append(vpnRoutes, prefix)
+			}
+
+			for _, route := range tt.localRoutes {
+				prefix, err := netip.ParsePrefix(route)
+				if err != nil {
+					t.Fatalf("Failed to parse local route %s: %v", route, err)
+				}
+				localRoutes = append(localRoutes, prefix)
+			}
+
+			isVpn, matchedPrefix := isVpnRoute(addr, vpnRoutes, localRoutes)
+			assert.Equal(t, tt.expectedVpn, isVpn, "isVpnRoute should return expectedVpn value")
+			assert.Equal(t, tt.expectedPrefix, matchedPrefix, "isVpnRoute should return expectedVpn prefix")
+		})
+	}
+}

client/internal/routemanager/systemops/systemops_linux_test.go

@@ -1,13 +1,10 @@
-//go:build !android
+//go:build linux && !android && privileged
 
 package systemops
 
 import (
 	"errors"
-	"fmt"
 	"net"
-	"os"
-	"strings"
 	"syscall"
 	"testing"
 
@@ -18,10 +15,6 @@ import (
 	"github.com/netbirdio/netbird/client/internal/routemanager/vars"
 )
 
-var expectedVPNint = "wgtest0"
-var expectedExternalInt = "dummyext0"
-var expectedInternalInt = "dummyint0"
-
 func init() {
 	testCases = append(testCases, []testCase{
 		{
@@ -33,62 +26,6 @@ func init() {
 	}...)
 }
 
-func TestEntryExists(t *testing.T) {
-	tempDir := t.TempDir()
-	tempFilePath := fmt.Sprintf("%s/rt_tables", tempDir)
-
-	content := []string{
-		"1000 reserved",
-		fmt.Sprintf("%d %s", NetbirdVPNTableID, NetbirdVPNTableName),
-		"9999 other_table",
-	}
-	require.NoError(t, os.WriteFile(tempFilePath, []byte(strings.Join(content, "\n")), 0644))
-
-	file, err := os.Open(tempFilePath)
-	require.NoError(t, err)
-	defer func() {
-		assert.NoError(t, file.Close())
-	}()
-
-	tests := []struct {
-		name        string
-		id          int
-		shouldExist bool
-		err         error
-	}{
-		{
-			name:        "ExistsWithNetbirdPrefix",
-			id:          7120,
-			shouldExist: true,
-			err:         nil,
-		},
-		{
-			name:        "ExistsWithDifferentName",
-			id:          1000,
-			shouldExist: true,
-			err:         ErrTableIDExists,
-		},
-		{
-			name:        "DoesNotExist",
-			id:          1234,
-			shouldExist: false,
-			err:         nil,
-		},
-	}
-
-	for _, tc := range tests {
-		t.Run(tc.name, func(t *testing.T) {
-			exists, err := entryExists(file, tc.id)
-			if tc.err != nil {
-				assert.ErrorIs(t, err, tc.err)
-			} else {
-				assert.NoError(t, err)
-			}
-			assert.Equal(t, tc.shouldExist, exists)
-		})
-	}
-}
-
 func createAndSetupDummyInterface(t *testing.T, interfaceName, ipAddressCIDR string) string {
 	t.Helper()
 

client/internal/routemanager/systemops/systemops_routing_data_linux_test.go

@@ -0,0 +1,15 @@
+//go:build linux && !android
+
+package systemops
+
+// Interface names used by the shared routing test fixtures. Kept untagged (no
+// privileged build tag) so the non-privileged test files in this package compile.
+//
+//nolint:unused // consumed by the privileged-tagged routing tests
+var expectedVPNint = "wgtest0"
+
+//nolint:unused // consumed by the privileged-tagged routing tests
+var expectedExternalInt = "dummyext0"
+
+//nolint:unused // consumed by the privileged-tagged routing tests
+var expectedInternalInt = "dummyint0"

client/internal/routemanager/systemops/systemops_routing_data_test.go

@@ -0,0 +1,83 @@
+//go:build (linux && !android) || (darwin && !ios) || freebsd || openbsd || netbsd || dragonfly
+
+package systemops
+
+import (
+	"net"
+
+	nbnet "github.com/netbirdio/netbird/client/net"
+)
+
+// Shared, non-privileged routing test fixtures. The privileged TestRouting (and its
+// per-platform init() appenders) consume these; they live here so the unprivileged
+// BSD/darwin test files compile without the privileged build tag.
+
+type PacketExpectation struct {
+	SrcIP   net.IP
+	DstIP   net.IP
+	SrcPort int
+	DstPort int
+	UDP     bool
+	TCP     bool
+}
+
+//nolint:unused // consumed by the privileged-tagged routing tests
+type testCase struct {
+	name              string
+	expectedInterface string
+	dialer            dialer
+	expectedPacket    PacketExpectation
+}
+
+//nolint:unused // consumed by the privileged-tagged routing tests
+var testCases = []testCase{
+	{
+		name:              "To external host without custom dialer via vpn",
+		expectedInterface: expectedVPNint,
+		dialer:            &net.Dialer{},
+		expectedPacket:    createPacketExpectation("100.64.0.1", 12345, "192.0.2.1", 53),
+	},
+	{
+		name:              "To external host with custom dialer via physical interface",
+		expectedInterface: expectedExternalInt,
+		dialer:            nbnet.NewDialer(),
+		expectedPacket:    createPacketExpectation("192.168.0.1", 12345, "192.0.2.1", 53),
+	},
+
+	{
+		name:              "To duplicate internal route with custom dialer via physical interface",
+		expectedInterface: expectedInternalInt,
+		dialer:            nbnet.NewDialer(),
+		expectedPacket:    createPacketExpectation("192.168.1.1", 12345, "10.0.0.2", 53),
+	},
+	{
+		name:              "To duplicate internal route without custom dialer via physical interface", // local route takes precedence
+		expectedInterface: expectedInternalInt,
+		dialer:            &net.Dialer{},
+		expectedPacket:    createPacketExpectation("192.168.1.1", 12345, "10.0.0.2", 53),
+	},
+
+	{
+		name:              "To unique vpn route with custom dialer via physical interface",
+		expectedInterface: expectedExternalInt,
+		dialer:            nbnet.NewDialer(),
+		expectedPacket:    createPacketExpectation("192.168.0.1", 12345, "172.16.0.2", 53),
+	},
+	{
+		name:              "To unique vpn route without custom dialer via vpn",
+		expectedInterface: expectedVPNint,
+		dialer:            &net.Dialer{},
+		expectedPacket:    createPacketExpectation("100.64.0.1", 12345, "172.16.0.2", 53),
+	},
+}
+
+//nolint:unused // consumed by the privileged-tagged routing tests
+func createPacketExpectation(srcIP string, srcPort int, dstIP string, dstPort int) PacketExpectation {
+	return PacketExpectation{
+		SrcIP:   net.ParseIP(srcIP),
+		DstIP:   net.ParseIP(dstIP),
+		SrcPort: srcPort,
+		DstPort: dstPort,
+		UDP:     true,
+	}
+}

client/internal/routemanager/systemops/systemops_unix_test.go

@@ -1,4 +1,4 @@
-//go:build (linux && !android) || (darwin && !ios) || freebsd || openbsd || netbsd || dragonfly
+//go:build ((linux && !android) || (darwin && !ios) || freebsd || openbsd || netbsd || dragonfly) && privileged
 
 package systemops
 
@@ -20,63 +20,6 @@ import (
 	nbnet "github.com/netbirdio/netbird/client/net"
 )
 
-type PacketExpectation struct {
-	SrcIP   net.IP
-	DstIP   net.IP
-	SrcPort int
-	DstPort int
-	UDP     bool
-	TCP     bool
-}
-
-type testCase struct {
-	name              string
-	expectedInterface string
-	dialer            dialer
-	expectedPacket    PacketExpectation
-}
-
-var testCases = []testCase{
-	{
-		name:              "To external host without custom dialer via vpn",
-		expectedInterface: expectedVPNint,
-		dialer:            &net.Dialer{},
-		expectedPacket:    createPacketExpectation("100.64.0.1", 12345, "192.0.2.1", 53),
-	},
-	{
-		name:              "To external host with custom dialer via physical interface",
-		expectedInterface: expectedExternalInt,
-		dialer:            nbnet.NewDialer(),
-		expectedPacket:    createPacketExpectation("192.168.0.1", 12345, "192.0.2.1", 53),
-	},
-
-	{
-		name:              "To duplicate internal route with custom dialer via physical interface",
-		expectedInterface: expectedInternalInt,
-		dialer:            nbnet.NewDialer(),
-		expectedPacket:    createPacketExpectation("192.168.1.1", 12345, "10.0.0.2", 53),
-	},
-	{
-		name:              "To duplicate internal route without custom dialer via physical interface", // local route takes precedence
-		expectedInterface: expectedInternalInt,
-		dialer:            &net.Dialer{},
-		expectedPacket:    createPacketExpectation("192.168.1.1", 12345, "10.0.0.2", 53),
-	},
-
-	{
-		name:              "To unique vpn route with custom dialer via physical interface",
-		expectedInterface: expectedExternalInt,
-		dialer:            nbnet.NewDialer(),
-		expectedPacket:    createPacketExpectation("192.168.0.1", 12345, "172.16.0.2", 53),
-	},
-	{
-		name:              "To unique vpn route without custom dialer via vpn",
-		expectedInterface: expectedVPNint,
-		dialer:            &net.Dialer{},
-		expectedPacket:    createPacketExpectation("100.64.0.1", 12345, "172.16.0.2", 53),
-	},
-}
-
 func TestRouting(t *testing.T) {
 	nbnet.Init()
 	for _, tc := range testCases {
@@ -102,16 +45,6 @@ func TestRouting(t *testing.T) {
 	}
 }
 
-func createPacketExpectation(srcIP string, srcPort int, dstIP string, dstPort int) PacketExpectation {
-	return PacketExpectation{
-		SrcIP:   net.ParseIP(srcIP),
-		DstIP:   net.ParseIP(dstIP),
-		SrcPort: srcPort,
-		DstPort: dstPort,
-		UDP:     true,
-	}
-}
-
 func startPacketCapture(t *testing.T, intf, filter string) *pcap.Handle {
 	t.Helper()
 

client/internal/routemanager/systemops/systemops_windows_test.go

@@ -1,3 +1,5 @@
+//go:build windows && privileged
+
 package systemops
 
 import (

client/internal/routemanager/systemops/v6route_bsd_test.go

@@ -11,6 +11,8 @@ import (
 // ensureIPv6DefaultRoute installs an IPv6 default route via the loopback
 // interface so route lookups for global IPv6 prefixes resolve in environments
 // without v6 connectivity. If a default already exists it is left alone.
+//
+//nolint:unused // consumed by the privileged-tagged routing tests
 func ensureIPv6DefaultRoute(t *testing.T) {
 	t.Helper()
 

client/internal/routemanager/systemops/v6route_linux_test.go

@@ -1,4 +1,4 @@
-//go:build linux && !android
+//go:build linux && !android && privileged
 
 package systemops
 

client/internal/routemanager/systemops/v6route_windows_test.go

@@ -8,11 +8,14 @@ import (
 	"testing"
 )
 
+//nolint:unused // consumed by the privileged-tagged routing tests
 const loopbackIfaceWindows = "Loopback Pseudo-Interface 1"
 
 // ensureIPv6DefaultRoute installs an IPv6 default route via the loopback
 // interface so route lookups for global IPv6 prefixes resolve in environments
 // without v6 connectivity. If a default already exists it is left alone.
+//
+//nolint:unused // consumed by the privileged-tagged routing tests
 func ensureIPv6DefaultRoute(t *testing.T) {
 	t.Helper()
 

client/server/server_privileged_test.go

@@ -0,0 +1,235 @@
+//go:build privileged
+
+package server
+
+import (
+	"context"
+	"net"
+	"os/user"
+	"testing"
+	"time"
+
+	"github.com/golang/mock/gomock"
+	"github.com/stretchr/testify/require"
+	"go.opentelemetry.io/otel"
+
+	"github.com/netbirdio/netbird/management/server/integrations/integrated_validator/validator"
+
+	"github.com/netbirdio/netbird/management/internals/controllers/network_map/controller"
+	"github.com/netbirdio/netbird/management/internals/controllers/network_map/update_channel"
+	"github.com/netbirdio/netbird/management/internals/modules/peers"
+	"github.com/netbirdio/netbird/management/internals/modules/peers/ephemeral/manager"
+	nbgrpc "github.com/netbirdio/netbird/management/internals/shared/grpc"
+	"github.com/netbirdio/netbird/management/server/job"
+
+	"github.com/netbirdio/netbird/management/internals/server/config"
+	"github.com/netbirdio/netbird/management/server/groups"
+
+	log "github.com/sirupsen/logrus"
+	"google.golang.org/grpc"
+	"google.golang.org/grpc/keepalive"
+
+	"github.com/netbirdio/netbird/client/internal"
+	"github.com/netbirdio/netbird/client/internal/peer"
+	"github.com/netbirdio/netbird/client/internal/profilemanager"
+	"github.com/netbirdio/netbird/management/server"
+	"github.com/netbirdio/netbird/management/server/activity"
+	nbcache "github.com/netbirdio/netbird/management/server/cache"
+	"github.com/netbirdio/netbird/management/server/integrations/port_forwarding"
+	"github.com/netbirdio/netbird/management/server/permissions"
+	"github.com/netbirdio/netbird/management/server/settings"
+	"github.com/netbirdio/netbird/management/server/store"
+	"github.com/netbirdio/netbird/management/server/telemetry"
+	mgmtProto "github.com/netbirdio/netbird/shared/management/proto"
+	"github.com/netbirdio/netbird/shared/signal/proto"
+	signalServer "github.com/netbirdio/netbird/signal/server"
+)
+
+var (
+	kaep = keepalive.EnforcementPolicy{
+		MinTime:             15 * time.Second,
+		PermitWithoutStream: true,
+	}
+
+	kasp = keepalive.ServerParameters{
+		MaxConnectionIdle:     15 * time.Second,
+		MaxConnectionAgeGrace: 5 * time.Second,
+		Time:                  5 * time.Second,
+		Timeout:               2 * time.Second,
+	}
+)
+
+// TestConnectWithRetryRuns checks that the connectWithRetry function runs and runs the retries according to the times specified via environment variables
+// we will use a management server started via to simulate the server and capture the number of retries
+func TestConnectWithRetryRuns(t *testing.T) {
+	// start the signal server
+	_, signalAddr, err := startSignal(t)
+	if err != nil {
+		t.Fatalf("failed to start signal server: %v", err)
+	}
+
+	counter := 0
+	// start the management server
+	_, mgmtAddr, err := startManagement(t, signalAddr, &counter)
+	if err != nil {
+		t.Fatalf("failed to start management server: %v", err)
+	}
+
+	ctx := internal.CtxInitState(context.Background())
+
+	ctx, cancel := context.WithDeadline(ctx, time.Now().Add(30*time.Second))
+	defer cancel()
+	// create new server
+	ic := profilemanager.ConfigInput{
+		ManagementURL: "http://" + mgmtAddr,
+		ConfigPath:    t.TempDir() + "/test-profile.json",
+	}
+
+	config, err := profilemanager.UpdateOrCreateConfig(ic)
+	if err != nil {
+		t.Fatalf("failed to create config: %v", err)
+	}
+
+	currUser, err := user.Current()
+	require.NoError(t, err)
+
+	pm := profilemanager.ServiceManager{}
+	err = pm.SetActiveProfileState(&profilemanager.ActiveProfileState{
+		ID:       "test-profile",
+		Username: currUser.Username,
+	})
+	if err != nil {
+		t.Fatalf("failed to set active profile state: %v", err)
+	}
+
+	s := New(ctx, "debug", "", false, false, false, false)
+
+	s.config = config
+
+	s.statusRecorder = peer.NewRecorder(config.ManagementURL.String())
+	t.Setenv(retryInitialIntervalVar, "1s")
+	t.Setenv(maxRetryIntervalVar, "2s")
+	t.Setenv(maxRetryTimeVar, "5s")
+	t.Setenv(retryMultiplierVar, "1")
+
+	s.connectWithRetryRuns(ctx, config, s.statusRecorder, nil, nil)
+	if counter < 3 {
+		t.Fatalf("expected counter > 2, got %d", counter)
+	}
+}
+
+type mockServer struct {
+	mgmtProto.ManagementServiceServer
+	counter *int
+}
+
+func (m *mockServer) Login(ctx context.Context, req *mgmtProto.EncryptedMessage) (*mgmtProto.EncryptedMessage, error) {
+	*m.counter++
+	return m.ManagementServiceServer.Login(ctx, req)
+}
+
+func startManagement(t *testing.T, signalAddr string, counter *int) (*grpc.Server, string, error) {
+	t.Helper()
+	dataDir := t.TempDir()
+
+	config := &config.Config{
+		Stuns:      []*config.Host{},
+		TURNConfig: &config.TURNConfig{},
+		Signal: &config.Host{
+			Proto: "http",
+			URI:   signalAddr,
+		},
+		Datadir:    dataDir,
+		HttpConfig: nil,
+	}
+
+	lis, err := net.Listen("tcp", "localhost:0")
+	if err != nil {
+		return nil, "", err
+	}
+	s := grpc.NewServer(grpc.KeepaliveEnforcementPolicy(kaep), grpc.KeepaliveParams(kasp))
+	store, cleanUp, err := store.NewTestStoreFromSQL(context.Background(), "", config.Datadir)
+	if err != nil {
+		return nil, "", err
+	}
+	t.Cleanup(cleanUp)
+
+	eventStore := &activity.InMemoryEventStore{}
+	if err != nil {
+		return nil, "", err
+	}
+
+	ctrl := gomock.NewController(t)
+	t.Cleanup(ctrl.Finish)
+
+	permissionsManagerMock := permissions.NewMockManager(ctrl)
+	peersManager := peers.NewManager(store, permissionsManagerMock)
+	settingsManagerMock := settings.NewMockManager(ctrl)
+
+	jobManager := job.NewJobManager(nil, store, peersManager)
+
+	cacheStore, err := nbcache.NewStore(context.Background(), 100*time.Millisecond, 300*time.Millisecond, 100)
+	if err != nil {
+		return nil, "", err
+	}
+
+	ia, _ := validator.NewIntegratedValidator(context.Background(), peersManager, settingsManagerMock, eventStore, cacheStore)
+
+	metrics, err := telemetry.NewDefaultAppMetrics(context.Background())
+	require.NoError(t, err)
+
+	settingsMockManager := settings.NewMockManager(ctrl)
+	groupsManager := groups.NewManagerMock()
+
+	requestBuffer := server.NewAccountRequestBuffer(context.Background(), store)
+	peersUpdateManager := update_channel.NewPeersUpdateManager(metrics)
+	networkMapController := controller.NewController(context.Background(), store, metrics, peersUpdateManager, requestBuffer, server.MockIntegratedValidator{}, settingsMockManager, "netbird.selfhosted", port_forwarding.NewControllerMock(), manager.NewEphemeralManager(store, peersManager), config)
+	accountManager, err := server.BuildManager(context.Background(), config, store, networkMapController, jobManager, nil, "", eventStore, nil, false, ia, metrics, port_forwarding.NewControllerMock(), settingsMockManager, permissionsManagerMock, false, cacheStore)
+	if err != nil {
+		return nil, "", err
+	}
+
+	secretsManager, err := nbgrpc.NewTimeBasedAuthSecretsManager(peersUpdateManager, config.TURNConfig, config.Relay, settingsMockManager, groupsManager)
+	if err != nil {
+		return nil, "", err
+	}
+	mgmtServer, err := nbgrpc.NewServer(config, accountManager, settingsMockManager, jobManager, secretsManager, nil, nil, &server.MockIntegratedValidator{}, networkMapController, nil, nil)
+	if err != nil {
+		return nil, "", err
+	}
+	mock := &mockServer{
+		ManagementServiceServer: mgmtServer,
+		counter:                 counter,
+	}
+	mgmtProto.RegisterManagementServiceServer(s, mock)
+	go func() {
+		if err = s.Serve(lis); err != nil {
+			log.Fatalf("failed to serve: %v", err)
+		}
+	}()
+
+	return s, lis.Addr().String(), nil
+}
+
+func startSignal(t *testing.T) (*grpc.Server, string, error) {
+	t.Helper()
+
+	s := grpc.NewServer(grpc.KeepaliveEnforcementPolicy(kaep), grpc.KeepaliveParams(kasp))
+
+	lis, err := net.Listen("tcp", "localhost:0")
+	if err != nil {
+		return nil, "", err
+	}
+
+	srv, err := signalServer.NewServer(context.Background(), otel.Meter(""))
+	require.NoError(t, err)
+	proto.RegisterSignalExchangeServer(s, srv)
+
+	go func() {
+		if err = s.Serve(lis); err != nil {
+			log.Fatalf("failed to serve: %v", err)
+		}
+	}()
+
+	return s, lis.Addr().String(), nil
+}

client/server/server_test.go

@@ -2,124 +2,22 @@ package server
 
 import (
 	"context"
-	"net"
 	"net/url"
 	"os/user"
 	"path/filepath"
 	"testing"
 	"time"
 
-	"github.com/golang/mock/gomock"
-	"github.com/stretchr/testify/require"
-	"go.opentelemetry.io/otel"
-
-	"github.com/netbirdio/netbird/management/server/integrations/integrated_validator/validator"
-
-	"github.com/netbirdio/netbird/management/internals/controllers/network_map/controller"
-	"github.com/netbirdio/netbird/management/internals/controllers/network_map/update_channel"
-	"github.com/netbirdio/netbird/management/internals/modules/peers"
-	"github.com/netbirdio/netbird/management/internals/modules/peers/ephemeral/manager"
-	nbgrpc "github.com/netbirdio/netbird/management/internals/shared/grpc"
-	"github.com/netbirdio/netbird/management/server/job"
-
-	"github.com/netbirdio/netbird/management/internals/server/config"
-	"github.com/netbirdio/netbird/management/server/groups"
-
 	log "github.com/sirupsen/logrus"
 	"github.com/stretchr/testify/assert"
+	"github.com/stretchr/testify/require"
 	"google.golang.org/grpc"
-	"google.golang.org/grpc/keepalive"
 
 	"github.com/netbirdio/netbird/client/internal"
-	"github.com/netbirdio/netbird/client/internal/peer"
 	"github.com/netbirdio/netbird/client/internal/profilemanager"
 	daemonProto "github.com/netbirdio/netbird/client/proto"
-	"github.com/netbirdio/netbird/management/server"
-	"github.com/netbirdio/netbird/management/server/activity"
-	nbcache "github.com/netbirdio/netbird/management/server/cache"
-	"github.com/netbirdio/netbird/management/server/integrations/port_forwarding"
-	"github.com/netbirdio/netbird/management/server/permissions"
-	"github.com/netbirdio/netbird/management/server/settings"
-	"github.com/netbirdio/netbird/management/server/store"
-	"github.com/netbirdio/netbird/management/server/telemetry"
-	mgmtProto "github.com/netbirdio/netbird/shared/management/proto"
-	"github.com/netbirdio/netbird/shared/signal/proto"
-	signalServer "github.com/netbirdio/netbird/signal/server"
 )
 
-var (
-	kaep = keepalive.EnforcementPolicy{
-		MinTime:             15 * time.Second,
-		PermitWithoutStream: true,
-	}
-
-	kasp = keepalive.ServerParameters{
-		MaxConnectionIdle:     15 * time.Second,
-		MaxConnectionAgeGrace: 5 * time.Second,
-		Time:                  5 * time.Second,
-		Timeout:               2 * time.Second,
-	}
-)
-
-// TestConnectWithRetryRuns checks that the connectWithRetry function runs and runs the retries according to the times specified via environment variables
-// we will use a management server started via to simulate the server and capture the number of retries
-func TestConnectWithRetryRuns(t *testing.T) {
-	// start the signal server
-	_, signalAddr, err := startSignal(t)
-	if err != nil {
-		t.Fatalf("failed to start signal server: %v", err)
-	}
-
-	counter := 0
-	// start the management server
-	_, mgmtAddr, err := startManagement(t, signalAddr, &counter)
-	if err != nil {
-		t.Fatalf("failed to start management server: %v", err)
-	}
-
-	ctx := internal.CtxInitState(context.Background())
-
-	ctx, cancel := context.WithDeadline(ctx, time.Now().Add(30*time.Second))
-	defer cancel()
-	// create new server
-	ic := profilemanager.ConfigInput{
-		ManagementURL: "http://" + mgmtAddr,
-		ConfigPath:    t.TempDir() + "/test-profile.json",
-	}
-
-	config, err := profilemanager.UpdateOrCreateConfig(ic)
-	if err != nil {
-		t.Fatalf("failed to create config: %v", err)
-	}
-
-	currUser, err := user.Current()
-	require.NoError(t, err)
-
-	pm := profilemanager.ServiceManager{}
-	err = pm.SetActiveProfileState(&profilemanager.ActiveProfileState{
-		ID:       "test-profile",
-		Username: currUser.Username,
-	})
-	if err != nil {
-		t.Fatalf("failed to set active profile state: %v", err)
-	}
-
-	s := New(ctx, "debug", "", false, false, false, false)
-
-	s.config = config
-
-	s.statusRecorder = peer.NewRecorder(config.ManagementURL.String())
-	t.Setenv(retryInitialIntervalVar, "1s")
-	t.Setenv(maxRetryIntervalVar, "2s")
-	t.Setenv(maxRetryTimeVar, "5s")
-	t.Setenv(retryMultiplierVar, "1")
-
-	s.connectWithRetryRuns(ctx, config, s.statusRecorder, nil, nil)
-	if counter < 3 {
-		t.Fatalf("expected counter > 2, got %d", counter)
-	}
-}
-
 func TestServer_Up(t *testing.T) {
 	tempDir := t.TempDir()
 	origDefaultProfileDir := profilemanager.DefaultConfigPathDir
@@ -259,119 +157,3 @@ func TestServer_SubcribeEvents(t *testing.T) {
 
 	assert.NoError(t, err)
 }
-
-type mockServer struct {
-	mgmtProto.ManagementServiceServer
-	counter *int
-}
-
-func (m *mockServer) Login(ctx context.Context, req *mgmtProto.EncryptedMessage) (*mgmtProto.EncryptedMessage, error) {
-	*m.counter++
-	return m.ManagementServiceServer.Login(ctx, req)
-}
-
-func startManagement(t *testing.T, signalAddr string, counter *int) (*grpc.Server, string, error) {
-	t.Helper()
-	dataDir := t.TempDir()
-
-	config := &config.Config{
-		Stuns:      []*config.Host{},
-		TURNConfig: &config.TURNConfig{},
-		Signal: &config.Host{
-			Proto: "http",
-			URI:   signalAddr,
-		},
-		Datadir:    dataDir,
-		HttpConfig: nil,
-	}
-
-	lis, err := net.Listen("tcp", "localhost:0")
-	if err != nil {
-		return nil, "", err
-	}
-	s := grpc.NewServer(grpc.KeepaliveEnforcementPolicy(kaep), grpc.KeepaliveParams(kasp))
-	store, cleanUp, err := store.NewTestStoreFromSQL(context.Background(), "", config.Datadir)
-	if err != nil {
-		return nil, "", err
-	}
-	t.Cleanup(cleanUp)
-
-	eventStore := &activity.InMemoryEventStore{}
-	if err != nil {
-		return nil, "", err
-	}
-
-	ctrl := gomock.NewController(t)
-	t.Cleanup(ctrl.Finish)
-
-	permissionsManagerMock := permissions.NewMockManager(ctrl)
-	peersManager := peers.NewManager(store, permissionsManagerMock)
-	settingsManagerMock := settings.NewMockManager(ctrl)
-
-	jobManager := job.NewJobManager(nil, store, peersManager)
-
-	cacheStore, err := nbcache.NewStore(context.Background(), 100*time.Millisecond, 300*time.Millisecond, 100)
-	if err != nil {
-		return nil, "", err
-	}
-
-	ia, _ := validator.NewIntegratedValidator(context.Background(), peersManager, settingsManagerMock, eventStore, cacheStore)
-
-	metrics, err := telemetry.NewDefaultAppMetrics(context.Background())
-	require.NoError(t, err)
-
-	settingsMockManager := settings.NewMockManager(ctrl)
-	groupsManager := groups.NewManagerMock()
-
-	requestBuffer := server.NewAccountRequestBuffer(context.Background(), store)
-	peersUpdateManager := update_channel.NewPeersUpdateManager(metrics)
-	networkMapController := controller.NewController(context.Background(), store, metrics, peersUpdateManager, requestBuffer, server.MockIntegratedValidator{}, settingsMockManager, "netbird.selfhosted", port_forwarding.NewControllerMock(), manager.NewEphemeralManager(store, peersManager), config)
-	accountManager, err := server.BuildManager(context.Background(), config, store, networkMapController, jobManager, nil, "", eventStore, nil, false, ia, metrics, port_forwarding.NewControllerMock(), settingsMockManager, permissionsManagerMock, false, cacheStore)
-	if err != nil {
-		return nil, "", err
-	}
-
-	secretsManager, err := nbgrpc.NewTimeBasedAuthSecretsManager(peersUpdateManager, config.TURNConfig, config.Relay, settingsMockManager, groupsManager)
-	if err != nil {
-		return nil, "", err
-	}
-	mgmtServer, err := nbgrpc.NewServer(config, accountManager, settingsMockManager, jobManager, secretsManager, nil, nil, &server.MockIntegratedValidator{}, networkMapController, nil, nil)
-	if err != nil {
-		return nil, "", err
-	}
-	mock := &mockServer{
-		ManagementServiceServer: mgmtServer,
-		counter:                 counter,
-	}
-	mgmtProto.RegisterManagementServiceServer(s, mock)
-	go func() {
-		if err = s.Serve(lis); err != nil {
-			log.Fatalf("failed to serve: %v", err)
-		}
-	}()
-
-	return s, lis.Addr().String(), nil
-}
-
-func startSignal(t *testing.T) (*grpc.Server, string, error) {
-	t.Helper()
-
-	s := grpc.NewServer(grpc.KeepaliveEnforcementPolicy(kaep), grpc.KeepaliveParams(kasp))
-
-	lis, err := net.Listen("tcp", "localhost:0")
-	if err != nil {
-		log.Fatalf("failed to listen: %v", err)
-	}
-
-	srv, err := signalServer.NewServer(context.Background(), otel.Meter(""))
-	require.NoError(t, err)
-	proto.RegisterSignalExchangeServer(s, srv)
-
-	go func() {
-		if err = s.Serve(lis); err != nil {
-			log.Fatalf("failed to serve: %v", err)
-		}
-	}()
-
-	return s, lis.Addr().String(), nil
-}

client/ssh/client/client_privileged_test.go

@@ -0,0 +1,118 @@
+//go:build privileged
+
+package client
+
+import (
+	"context"
+	"errors"
+	"runtime"
+	"strings"
+	"testing"
+	"time"
+
+	"github.com/stretchr/testify/assert"
+	"github.com/stretchr/testify/require"
+	cryptossh "golang.org/x/crypto/ssh"
+
+	"github.com/netbirdio/netbird/client/ssh/testutil"
+)
+
+func TestSSHClient_CommandExecution(t *testing.T) {
+	if runtime.GOOS == "windows" && testutil.IsCI() {
+		t.Skip("Skipping Windows command execution tests in CI due to S4U authentication issues")
+	}
+
+	server, _, client := setupTestSSHServerAndClient(t)
+	defer func() {
+		err := server.Stop()
+		require.NoError(t, err)
+	}()
+	defer func() {
+		err := client.Close()
+		assert.NoError(t, err)
+	}()
+
+	ctx, cancel := context.WithTimeout(context.Background(), 3*time.Second)
+	defer cancel()
+
+	t.Run("ExecuteCommand captures output", func(t *testing.T) {
+		output, err := client.ExecuteCommand(ctx, "echo hello")
+		assert.NoError(t, err)
+		assert.Contains(t, string(output), "hello")
+	})
+
+	t.Run("ExecuteCommandWithIO streams output", func(t *testing.T) {
+		err := client.ExecuteCommandWithIO(ctx, "echo world")
+		assert.NoError(t, err)
+	})
+
+	t.Run("commands with flags work", func(t *testing.T) {
+		output, err := client.ExecuteCommand(ctx, "echo -n test_flag")
+		assert.NoError(t, err)
+		assert.Equal(t, "test_flag", strings.TrimSpace(string(output)))
+	})
+
+	t.Run("non-zero exit codes don't return errors", func(t *testing.T) {
+		var testCmd string
+		if runtime.GOOS == "windows" {
+			testCmd = "echo hello | Select-String notfound"
+		} else {
+			testCmd = "echo 'hello' | grep 'notfound'"
+		}
+		_, err := client.ExecuteCommand(ctx, testCmd)
+		assert.NoError(t, err)
+	})
+}
+
+func TestSSHClient_ContextCancellation(t *testing.T) {
+	server, serverAddr, _ := setupTestSSHServerAndClient(t)
+	defer func() {
+		err := server.Stop()
+		require.NoError(t, err)
+	}()
+
+	t.Run("connection with short timeout", func(t *testing.T) {
+		ctx, cancel := context.WithTimeout(context.Background(), 1*time.Millisecond)
+		defer cancel()
+
+		currentUser := testutil.GetTestUsername(t)
+		_, err := Dial(ctx, serverAddr, currentUser, DialOptions{
+			InsecureSkipVerify: true,
+		})
+		if err != nil {
+			// Check for actual timeout-related errors rather than string matching
+			assert.True(t,
+				errors.Is(err, context.DeadlineExceeded) ||
+					errors.Is(err, context.Canceled) ||
+					strings.Contains(err.Error(), "timeout"),
+				"Expected timeout-related error, got: %v", err)
+		}
+	})
+
+	t.Run("command execution cancellation", func(t *testing.T) {
+		ctx, cancel := context.WithTimeout(context.Background(), 5*time.Second)
+		defer cancel()
+		currentUser := testutil.GetTestUsername(t)
+		client, err := Dial(ctx, serverAddr, currentUser, DialOptions{
+			InsecureSkipVerify: true,
+		})
+		require.NoError(t, err)
+		defer func() {
+			if err := client.Close(); err != nil {
+				t.Logf("client close error: %v", err)
+			}
+		}()
+
+		cmdCtx, cmdCancel := context.WithTimeout(context.Background(), 100*time.Millisecond)
+		defer cmdCancel()
+
+		err = client.ExecuteCommandWithPTY(cmdCtx, "sleep 10")
+		if err != nil {
+			var exitMissingErr *cryptossh.ExitMissingError
+			isValidCancellation := errors.Is(err, context.DeadlineExceeded) ||
+				errors.Is(err, context.Canceled) ||
+				errors.As(err, &exitMissingErr)
+			assert.True(t, isValidCancellation, "Should handle command cancellation properly")
+		}
+	})
+}

client/ssh/client/client_test.go

@@ -15,7 +15,6 @@ import (
 
 	"github.com/stretchr/testify/assert"
 	"github.com/stretchr/testify/require"
-	cryptossh "golang.org/x/crypto/ssh"
 
 	"github.com/netbirdio/netbird/client/ssh"
 	sshserver "github.com/netbirdio/netbird/client/ssh/server"
@@ -78,53 +77,6 @@ func TestSSHClient_DialWithKey(t *testing.T) {
 	assert.NotNil(t, client.client)
 }
 
-func TestSSHClient_CommandExecution(t *testing.T) {
-	if runtime.GOOS == "windows" && testutil.IsCI() {
-		t.Skip("Skipping Windows command execution tests in CI due to S4U authentication issues")
-	}
-
-	server, _, client := setupTestSSHServerAndClient(t)
-	defer func() {
-		err := server.Stop()
-		require.NoError(t, err)
-	}()
-	defer func() {
-		err := client.Close()
-		assert.NoError(t, err)
-	}()
-
-	ctx, cancel := context.WithTimeout(context.Background(), 3*time.Second)
-	defer cancel()
-
-	t.Run("ExecuteCommand captures output", func(t *testing.T) {
-		output, err := client.ExecuteCommand(ctx, "echo hello")
-		assert.NoError(t, err)
-		assert.Contains(t, string(output), "hello")
-	})
-
-	t.Run("ExecuteCommandWithIO streams output", func(t *testing.T) {
-		err := client.ExecuteCommandWithIO(ctx, "echo world")
-		assert.NoError(t, err)
-	})
-
-	t.Run("commands with flags work", func(t *testing.T) {
-		output, err := client.ExecuteCommand(ctx, "echo -n test_flag")
-		assert.NoError(t, err)
-		assert.Equal(t, "test_flag", strings.TrimSpace(string(output)))
-	})
-
-	t.Run("non-zero exit codes don't return errors", func(t *testing.T) {
-		var testCmd string
-		if runtime.GOOS == "windows" {
-			testCmd = "echo hello | Select-String notfound"
-		} else {
-			testCmd = "echo 'hello' | grep 'notfound'"
-		}
-		_, err := client.ExecuteCommand(ctx, testCmd)
-		assert.NoError(t, err)
-	})
-}
-
 func TestSSHClient_ConnectionHandling(t *testing.T) {
 	server, serverAddr, _ := setupTestSSHServerAndClient(t)
 	defer func() {
@@ -154,59 +106,6 @@ func TestSSHClient_ConnectionHandling(t *testing.T) {
 	}
 }
 
-func TestSSHClient_ContextCancellation(t *testing.T) {
-	server, serverAddr, _ := setupTestSSHServerAndClient(t)
-	defer func() {
-		err := server.Stop()
-		require.NoError(t, err)
-	}()
-
-	t.Run("connection with short timeout", func(t *testing.T) {
-		ctx, cancel := context.WithTimeout(context.Background(), 1*time.Millisecond)
-		defer cancel()
-
-		currentUser := testutil.GetTestUsername(t)
-		_, err := Dial(ctx, serverAddr, currentUser, DialOptions{
-			InsecureSkipVerify: true,
-		})
-		if err != nil {
-			// Check for actual timeout-related errors rather than string matching
-			assert.True(t,
-				errors.Is(err, context.DeadlineExceeded) ||
-					errors.Is(err, context.Canceled) ||
-					strings.Contains(err.Error(), "timeout"),
-				"Expected timeout-related error, got: %v", err)
-		}
-	})
-
-	t.Run("command execution cancellation", func(t *testing.T) {
-		ctx, cancel := context.WithTimeout(context.Background(), 5*time.Second)
-		defer cancel()
-		currentUser := testutil.GetTestUsername(t)
-		client, err := Dial(ctx, serverAddr, currentUser, DialOptions{
-			InsecureSkipVerify: true,
-		})
-		require.NoError(t, err)
-		defer func() {
-			if err := client.Close(); err != nil {
-				t.Logf("client close error: %v", err)
-			}
-		}()
-
-		cmdCtx, cmdCancel := context.WithTimeout(context.Background(), 100*time.Millisecond)
-		defer cmdCancel()
-
-		err = client.ExecuteCommandWithPTY(cmdCtx, "sleep 10")
-		if err != nil {
-			var exitMissingErr *cryptossh.ExitMissingError
-			isValidCancellation := errors.Is(err, context.DeadlineExceeded) ||
-				errors.Is(err, context.Canceled) ||
-				errors.As(err, &exitMissingErr)
-			assert.True(t, isValidCancellation, "Should handle command cancellation properly")
-		}
-	})
-}
-
 func TestSSHClient_NoAuthMode(t *testing.T) {
 	hostKey, err := ssh.GeneratePrivateKey(ssh.ED25519)
 	require.NoError(t, err)

client/ssh/proxy/proxy_privileged_test.go

@@ -0,0 +1,423 @@
+//go:build privileged
+
+package proxy
+
+import (
+	"bytes"
+	"context"
+	"crypto/rand"
+	"crypto/rsa"
+	"encoding/base64"
+	"encoding/json"
+	"io"
+	"math/big"
+	"net"
+	"net/http"
+	"net/http/httptest"
+	"os"
+	"runtime"
+	"strconv"
+	"testing"
+	"time"
+
+	"github.com/golang-jwt/jwt/v5"
+	"github.com/stretchr/testify/assert"
+	"github.com/stretchr/testify/require"
+	cryptossh "golang.org/x/crypto/ssh"
+
+	nbssh "github.com/netbirdio/netbird/client/ssh"
+	sshauth "github.com/netbirdio/netbird/client/ssh/auth"
+	"github.com/netbirdio/netbird/client/ssh/server"
+	"github.com/netbirdio/netbird/client/ssh/testutil"
+	nbjwt "github.com/netbirdio/netbird/shared/auth/jwt"
+	sshuserhash "github.com/netbirdio/netbird/shared/sshauth"
+)
+
+func (m *mockDaemon) setJWTToken(token string) {
+	m.impl.jwtToken = token
+}
+
+func TestSSHProxy_Connect(t *testing.T) {
+	if testing.Short() {
+		t.Skip("Skipping integration test in short mode")
+	}
+
+	// TODO: Windows test times out - user switching and command execution tested on Linux
+	if runtime.GOOS == "windows" {
+		t.Skip("Skipping on Windows - covered by Linux tests")
+	}
+
+	const (
+		issuer   = "https://test-issuer.example.com"
+		audience = "test-audience"
+	)
+
+	jwksServer, privateKey, jwksURL := setupJWKSServer(t)
+	defer jwksServer.Close()
+
+	hostKey, err := nbssh.GeneratePrivateKey(nbssh.ED25519)
+	require.NoError(t, err)
+	hostPubKey, err := nbssh.GeneratePublicKey(hostKey)
+	require.NoError(t, err)
+
+	serverConfig := &server.Config{
+		HostKeyPEM: hostKey,
+		JWT: &server.JWTConfig{
+			Issuer:       issuer,
+			Audiences:    []string{audience},
+			KeysLocation: jwksURL,
+		},
+	}
+	sshServer := server.New(serverConfig)
+	sshServer.SetAllowRootLogin(true)
+
+	// Configure SSH authorization for the test user
+	testUsername := testutil.GetTestUsername(t)
+	testJWTUser := "test-username"
+	testUserHash, err := sshuserhash.HashUserID(testJWTUser)
+	require.NoError(t, err)
+
+	authConfig := &sshauth.Config{
+		UserIDClaim:     sshauth.DefaultUserIDClaim,
+		AuthorizedUsers: []sshuserhash.UserIDHash{testUserHash},
+		MachineUsers: map[string][]uint32{
+			testUsername: {0}, // Index 0 in AuthorizedUsers
+		},
+	}
+	sshServer.UpdateSSHAuth(authConfig)
+
+	sshServerAddr := server.StartTestServer(t, sshServer)
+	defer func() { _ = sshServer.Stop() }()
+
+	mockDaemon := startMockDaemon(t)
+	defer mockDaemon.stop()
+
+	host, portStr, err := net.SplitHostPort(sshServerAddr)
+	require.NoError(t, err)
+	port, err := strconv.Atoi(portStr)
+	require.NoError(t, err)
+
+	mockDaemon.setHostKey(host, hostPubKey)
+
+	validToken := generateValidJWT(t, privateKey, issuer, audience, testJWTUser)
+	mockDaemon.setJWTToken(validToken)
+
+	proxyInstance, err := New(mockDaemon.addr, host, port, io.Discard, nil)
+	require.NoError(t, err)
+
+	clientConn, proxyConn := net.Pipe()
+	defer func() { _ = clientConn.Close() }()
+
+	origStdin := os.Stdin
+	origStdout := os.Stdout
+	defer func() {
+		os.Stdin = origStdin
+		os.Stdout = origStdout
+	}()
+
+	stdinReader, stdinWriter, err := os.Pipe()
+	require.NoError(t, err)
+	stdoutReader, stdoutWriter, err := os.Pipe()
+	require.NoError(t, err)
+
+	os.Stdin = stdinReader
+	os.Stdout = stdoutWriter
+
+	go func() {
+		_, _ = io.Copy(stdinWriter, proxyConn)
+	}()
+	go func() {
+		_, _ = io.Copy(proxyConn, stdoutReader)
+	}()
+
+	ctx, cancel := context.WithTimeout(context.Background(), 5*time.Second)
+	defer cancel()
+
+	connectErrCh := make(chan error, 1)
+	go func() {
+		connectErrCh <- proxyInstance.Connect(ctx)
+	}()
+
+	sshConfig := &cryptossh.ClientConfig{
+		User:            testutil.GetTestUsername(t),
+		Auth:            []cryptossh.AuthMethod{},
+		HostKeyCallback: cryptossh.InsecureIgnoreHostKey(),
+		Timeout:         3 * time.Second,
+	}
+
+	sshClientConn, chans, reqs, err := cryptossh.NewClientConn(clientConn, "test", sshConfig)
+	require.NoError(t, err, "Should connect to proxy server")
+	defer func() { _ = sshClientConn.Close() }()
+
+	sshClient := cryptossh.NewClient(sshClientConn, chans, reqs)
+
+	session, err := sshClient.NewSession()
+	require.NoError(t, err, "Should create session through full proxy to backend")
+
+	outputCh := make(chan []byte, 1)
+	errCh := make(chan error, 1)
+	go func() {
+		output, err := session.Output("echo hello-from-proxy")
+		outputCh <- output
+		errCh <- err
+	}()
+
+	select {
+	case output := <-outputCh:
+		err := <-errCh
+		require.NoError(t, err, "Command should execute successfully through proxy")
+		assert.Contains(t, string(output), "hello-from-proxy", "Should receive command output through proxy")
+	case <-time.After(3 * time.Second):
+		t.Fatal("Command execution timed out")
+	}
+
+	_ = session.Close()
+	_ = sshClient.Close()
+	_ = clientConn.Close()
+	cancel()
+}
+
+// TestSSHProxy_CommandQuoting verifies that the proxy preserves shell quoting
+// when forwarding commands to the backend. This is critical for tools like
+// Ansible that send commands such as:
+//
+//	/bin/sh -c '( umask 77 && mkdir -p ... ) && sleep 0'
+//
+// The single quotes must be preserved so the backend shell receives the
+// subshell expression as a single argument to -c.
+func TestSSHProxy_CommandQuoting(t *testing.T) {
+	if testing.Short() {
+		t.Skip("Skipping integration test in short mode")
+	}
+
+	sshClient, cleanup := setupProxySSHClient(t)
+	defer cleanup()
+
+	// These commands simulate what the SSH protocol delivers as exec payloads.
+	// When a user types: ssh host '/bin/sh -c "( echo hello )"'
+	// the local shell strips the outer single quotes, and the SSH exec request
+	// contains the raw string: /bin/sh -c "( echo hello )"
+	//
+	// The proxy must forward this string verbatim. Using session.Command()
+	// (shlex.Split + strings.Join) strips the inner double quotes, breaking
+	// the command on the backend.
+	tests := []struct {
+		name    string
+		command string
+		expect  string
+	}{
+		{
+			name:    "subshell_in_double_quotes",
+			command: `/bin/sh -c "( echo from-subshell ) && echo outer"`,
+			expect:  "from-subshell\nouter\n",
+		},
+		{
+			name:    "printf_with_special_chars",
+			command: `/bin/sh -c "printf '%s\n' 'hello world'"`,
+			expect:  "hello world\n",
+		},
+		{
+			name:    "nested_command_substitution",
+			command: `/bin/sh -c "echo $(echo nested)"`,
+			expect:  "nested\n",
+		},
+	}
+
+	for _, tc := range tests {
+		t.Run(tc.name, func(t *testing.T) {
+			session, err := sshClient.NewSession()
+			require.NoError(t, err)
+			defer func() { _ = session.Close() }()
+
+			var stderrBuf bytes.Buffer
+			session.Stderr = &stderrBuf
+
+			outputCh := make(chan []byte, 1)
+			errCh := make(chan error, 1)
+			go func() {
+				output, err := session.Output(tc.command)
+				outputCh <- output
+				errCh <- err
+			}()
+
+			select {
+			case output := <-outputCh:
+				err := <-errCh
+				if stderrBuf.Len() > 0 {
+					t.Logf("stderr: %s", stderrBuf.String())
+				}
+				require.NoError(t, err, "command should succeed: %s", tc.command)
+				assert.Equal(t, tc.expect, string(output), "output mismatch for: %s", tc.command)
+			case <-time.After(5 * time.Second):
+				t.Fatalf("command timed out: %s", tc.command)
+			}
+		})
+	}
+}
+
+// setupProxySSHClient creates a full proxy test environment and returns
+// an SSH client connected through the proxy to a backend NetBird SSH server.
+func setupProxySSHClient(t *testing.T) (*cryptossh.Client, func()) {
+	t.Helper()
+
+	const (
+		issuer   = "https://test-issuer.example.com"
+		audience = "test-audience"
+	)
+
+	jwksServer, privateKey, jwksURL := setupJWKSServer(t)
+
+	hostKey, err := nbssh.GeneratePrivateKey(nbssh.ED25519)
+	require.NoError(t, err)
+	hostPubKey, err := nbssh.GeneratePublicKey(hostKey)
+	require.NoError(t, err)
+
+	serverConfig := &server.Config{
+		HostKeyPEM: hostKey,
+		JWT: &server.JWTConfig{
+			Issuer:       issuer,
+			Audiences:    []string{audience},
+			KeysLocation: jwksURL,
+		},
+	}
+	sshServer := server.New(serverConfig)
+	sshServer.SetAllowRootLogin(true)
+
+	testUsername := testutil.GetTestUsername(t)
+	testJWTUser := "test-username"
+	testUserHash, err := sshuserhash.HashUserID(testJWTUser)
+	require.NoError(t, err)
+
+	authConfig := &sshauth.Config{
+		UserIDClaim:     sshauth.DefaultUserIDClaim,
+		AuthorizedUsers: []sshuserhash.UserIDHash{testUserHash},
+		MachineUsers: map[string][]uint32{
+			testUsername: {0},
+		},
+	}
+	sshServer.UpdateSSHAuth(authConfig)
+
+	sshServerAddr := server.StartTestServer(t, sshServer)
+
+	mockDaemon := startMockDaemon(t)
+
+	host, portStr, err := net.SplitHostPort(sshServerAddr)
+	require.NoError(t, err)
+	port, err := strconv.Atoi(portStr)
+	require.NoError(t, err)
+
+	mockDaemon.setHostKey(host, hostPubKey)
+
+	validToken := generateValidJWT(t, privateKey, issuer, audience, testJWTUser)
+	mockDaemon.setJWTToken(validToken)
+
+	proxyInstance, err := New(mockDaemon.addr, host, port, io.Discard, nil)
+	require.NoError(t, err)
+
+	origStdin := os.Stdin
+	origStdout := os.Stdout
+
+	stdinReader, stdinWriter, err := os.Pipe()
+	require.NoError(t, err)
+	stdoutReader, stdoutWriter, err := os.Pipe()
+	require.NoError(t, err)
+
+	os.Stdin = stdinReader
+	os.Stdout = stdoutWriter
+
+	clientConn, proxyConn := net.Pipe()
+
+	go func() { _, _ = io.Copy(stdinWriter, proxyConn) }()
+	go func() { _, _ = io.Copy(proxyConn, stdoutReader) }()
+
+	ctx, cancel := context.WithTimeout(context.Background(), 10*time.Second)
+
+	go func() {
+		_ = proxyInstance.Connect(ctx)
+	}()
+
+	sshConfig := &cryptossh.ClientConfig{
+		User:            testutil.GetTestUsername(t),
+		Auth:            []cryptossh.AuthMethod{},
+		HostKeyCallback: cryptossh.InsecureIgnoreHostKey(),
+		Timeout:         5 * time.Second,
+	}
+
+	sshClientConn, chans, reqs, err := cryptossh.NewClientConn(clientConn, "test", sshConfig)
+	require.NoError(t, err)
+
+	client := cryptossh.NewClient(sshClientConn, chans, reqs)
+
+	cleanupFn := func() {
+		_ = client.Close()
+		_ = clientConn.Close()
+		cancel()
+		os.Stdin = origStdin
+		os.Stdout = origStdout
+		_ = sshServer.Stop()
+		mockDaemon.stop()
+		jwksServer.Close()
+	}
+
+	return client, cleanupFn
+}
+
+func setupJWKSServer(t *testing.T) (*httptest.Server, *rsa.PrivateKey, string) {
+	t.Helper()
+	privateKey, jwksJSON := generateTestJWKS(t)
+
+	server := httptest.NewServer(http.HandlerFunc(func(w http.ResponseWriter, r *http.Request) {
+		w.Header().Set("Content-Type", "application/json")
+		if _, err := w.Write(jwksJSON); err != nil {
+			http.Error(w, err.Error(), http.StatusInternalServerError)
+		}
+	}))
+
+	return server, privateKey, server.URL
+}
+
+func generateTestJWKS(t *testing.T) (*rsa.PrivateKey, []byte) {
+	t.Helper()
+	privateKey, err := rsa.GenerateKey(rand.Reader, 2048)
+	require.NoError(t, err)
+
+	publicKey := &privateKey.PublicKey
+	n := publicKey.N.Bytes()
+	e := publicKey.E
+
+	jwk := nbjwt.JSONWebKey{
+		Kty: "RSA",
+		Kid: "test-key-id",
+		Use: "sig",
+		N:   base64.RawURLEncoding.EncodeToString(n),
+		E:   base64.RawURLEncoding.EncodeToString(big.NewInt(int64(e)).Bytes()),
+	}
+
+	jwks := nbjwt.Jwks{
+		Keys: []nbjwt.JSONWebKey{jwk},
+	}
+
+	jwksJSON, err := json.Marshal(jwks)
+	require.NoError(t, err)
+
+	return privateKey, jwksJSON
+}
+
+func generateValidJWT(t *testing.T, privateKey *rsa.PrivateKey, issuer, audience string, user string) string {
+	t.Helper()
+	claims := jwt.MapClaims{
+		"iss": issuer,
+		"aud": audience,
+		"sub": user,
+		"exp": time.Now().Add(time.Hour).Unix(),
+		"iat": time.Now().Unix(),
+	}
+
+	token := jwt.NewWithClaims(jwt.SigningMethodRS256, claims)
+	token.Header["kid"] = "test-key-id"
+
+	tokenString, err := token.SignedString(privateKey)
+	require.NoError(t, err)
+
+	return tokenString
+}

client/ssh/proxy/proxy_test.go

@@ -1,25 +1,12 @@
 package proxy
 
 import (
-	"bytes"
 	"context"
-	"crypto/rand"
-	"crypto/rsa"
-	"encoding/base64"
-	"encoding/json"
 	"fmt"
-	"io"
-	"math/big"
 	"net"
-	"net/http"
-	"net/http/httptest"
 	"os"
-	"runtime"
-	"strconv"
 	"testing"
-	"time"
 
-	"github.com/golang-jwt/jwt/v5"
 	"github.com/stretchr/testify/assert"
 	"github.com/stretchr/testify/require"
 	cryptossh "golang.org/x/crypto/ssh"
@@ -28,11 +15,7 @@ import (
 
 	"github.com/netbirdio/netbird/client/proto"
 	nbssh "github.com/netbirdio/netbird/client/ssh"
-	sshauth "github.com/netbirdio/netbird/client/ssh/auth"
-	"github.com/netbirdio/netbird/client/ssh/server"
 	"github.com/netbirdio/netbird/client/ssh/testutil"
-	nbjwt "github.com/netbirdio/netbird/shared/auth/jwt"
-	sshuserhash "github.com/netbirdio/netbird/shared/sshauth"
 )
 
 func TestMain(m *testing.M) {
@@ -106,331 +89,6 @@ func TestSSHProxy_verifyHostKey(t *testing.T) {
 	})
 }
 
-func TestSSHProxy_Connect(t *testing.T) {
-	if testing.Short() {
-		t.Skip("Skipping integration test in short mode")
-	}
-
-	// TODO: Windows test times out - user switching and command execution tested on Linux
-	if runtime.GOOS == "windows" {
-		t.Skip("Skipping on Windows - covered by Linux tests")
-	}
-
-	const (
-		issuer   = "https://test-issuer.example.com"
-		audience = "test-audience"
-	)
-
-	jwksServer, privateKey, jwksURL := setupJWKSServer(t)
-	defer jwksServer.Close()
-
-	hostKey, err := nbssh.GeneratePrivateKey(nbssh.ED25519)
-	require.NoError(t, err)
-	hostPubKey, err := nbssh.GeneratePublicKey(hostKey)
-	require.NoError(t, err)
-
-	serverConfig := &server.Config{
-		HostKeyPEM: hostKey,
-		JWT: &server.JWTConfig{
-			Issuer:       issuer,
-			Audiences:    []string{audience},
-			KeysLocation: jwksURL,
-		},
-	}
-	sshServer := server.New(serverConfig)
-	sshServer.SetAllowRootLogin(true)
-
-	// Configure SSH authorization for the test user
-	testUsername := testutil.GetTestUsername(t)
-	testJWTUser := "test-username"
-	testUserHash, err := sshuserhash.HashUserID(testJWTUser)
-	require.NoError(t, err)
-
-	authConfig := &sshauth.Config{
-		UserIDClaim:     sshauth.DefaultUserIDClaim,
-		AuthorizedUsers: []sshuserhash.UserIDHash{testUserHash},
-		MachineUsers: map[string][]uint32{
-			testUsername: {0}, // Index 0 in AuthorizedUsers
-		},
-	}
-	sshServer.UpdateSSHAuth(authConfig)
-
-	sshServerAddr := server.StartTestServer(t, sshServer)
-	defer func() { _ = sshServer.Stop() }()
-
-	mockDaemon := startMockDaemon(t)
-	defer mockDaemon.stop()
-
-	host, portStr, err := net.SplitHostPort(sshServerAddr)
-	require.NoError(t, err)
-	port, err := strconv.Atoi(portStr)
-	require.NoError(t, err)
-
-	mockDaemon.setHostKey(host, hostPubKey)
-
-	validToken := generateValidJWT(t, privateKey, issuer, audience, testJWTUser)
-	mockDaemon.setJWTToken(validToken)
-
-	proxyInstance, err := New(mockDaemon.addr, host, port, io.Discard, nil)
-	require.NoError(t, err)
-
-	clientConn, proxyConn := net.Pipe()
-	defer func() { _ = clientConn.Close() }()
-
-	origStdin := os.Stdin
-	origStdout := os.Stdout
-	defer func() {
-		os.Stdin = origStdin
-		os.Stdout = origStdout
-	}()
-
-	stdinReader, stdinWriter, err := os.Pipe()
-	require.NoError(t, err)
-	stdoutReader, stdoutWriter, err := os.Pipe()
-	require.NoError(t, err)
-
-	os.Stdin = stdinReader
-	os.Stdout = stdoutWriter
-
-	go func() {
-		_, _ = io.Copy(stdinWriter, proxyConn)
-	}()
-	go func() {
-		_, _ = io.Copy(proxyConn, stdoutReader)
-	}()
-
-	ctx, cancel := context.WithTimeout(context.Background(), 5*time.Second)
-	defer cancel()
-
-	connectErrCh := make(chan error, 1)
-	go func() {
-		connectErrCh <- proxyInstance.Connect(ctx)
-	}()
-
-	sshConfig := &cryptossh.ClientConfig{
-		User:            testutil.GetTestUsername(t),
-		Auth:            []cryptossh.AuthMethod{},
-		HostKeyCallback: cryptossh.InsecureIgnoreHostKey(),
-		Timeout:         3 * time.Second,
-	}
-
-	sshClientConn, chans, reqs, err := cryptossh.NewClientConn(clientConn, "test", sshConfig)
-	require.NoError(t, err, "Should connect to proxy server")
-	defer func() { _ = sshClientConn.Close() }()
-
-	sshClient := cryptossh.NewClient(sshClientConn, chans, reqs)
-
-	session, err := sshClient.NewSession()
-	require.NoError(t, err, "Should create session through full proxy to backend")
-
-	outputCh := make(chan []byte, 1)
-	errCh := make(chan error, 1)
-	go func() {
-		output, err := session.Output("echo hello-from-proxy")
-		outputCh <- output
-		errCh <- err
-	}()
-
-	select {
-	case output := <-outputCh:
-		err := <-errCh
-		require.NoError(t, err, "Command should execute successfully through proxy")
-		assert.Contains(t, string(output), "hello-from-proxy", "Should receive command output through proxy")
-	case <-time.After(3 * time.Second):
-		t.Fatal("Command execution timed out")
-	}
-
-	_ = session.Close()
-	_ = sshClient.Close()
-	_ = clientConn.Close()
-	cancel()
-}
-
-// TestSSHProxy_CommandQuoting verifies that the proxy preserves shell quoting
-// when forwarding commands to the backend. This is critical for tools like
-// Ansible that send commands such as:
-//
-//	/bin/sh -c '( umask 77 && mkdir -p ... ) && sleep 0'
-//
-// The single quotes must be preserved so the backend shell receives the
-// subshell expression as a single argument to -c.
-func TestSSHProxy_CommandQuoting(t *testing.T) {
-	if testing.Short() {
-		t.Skip("Skipping integration test in short mode")
-	}
-
-	sshClient, cleanup := setupProxySSHClient(t)
-	defer cleanup()
-
-	// These commands simulate what the SSH protocol delivers as exec payloads.
-	// When a user types: ssh host '/bin/sh -c "( echo hello )"'
-	// the local shell strips the outer single quotes, and the SSH exec request
-	// contains the raw string: /bin/sh -c "( echo hello )"
-	//
-	// The proxy must forward this string verbatim. Using session.Command()
-	// (shlex.Split + strings.Join) strips the inner double quotes, breaking
-	// the command on the backend.
-	tests := []struct {
-		name    string
-		command string
-		expect  string
-	}{
-		{
-			name:    "subshell_in_double_quotes",
-			command: `/bin/sh -c "( echo from-subshell ) && echo outer"`,
-			expect:  "from-subshell\nouter\n",
-		},
-		{
-			name:    "printf_with_special_chars",
-			command: `/bin/sh -c "printf '%s\n' 'hello world'"`,
-			expect:  "hello world\n",
-		},
-		{
-			name:    "nested_command_substitution",
-			command: `/bin/sh -c "echo $(echo nested)"`,
-			expect:  "nested\n",
-		},
-	}
-
-	for _, tc := range tests {
-		t.Run(tc.name, func(t *testing.T) {
-			session, err := sshClient.NewSession()
-			require.NoError(t, err)
-			defer func() { _ = session.Close() }()
-
-			var stderrBuf bytes.Buffer
-			session.Stderr = &stderrBuf
-
-			outputCh := make(chan []byte, 1)
-			errCh := make(chan error, 1)
-			go func() {
-				output, err := session.Output(tc.command)
-				outputCh <- output
-				errCh <- err
-			}()
-
-			select {
-			case output := <-outputCh:
-				err := <-errCh
-				if stderrBuf.Len() > 0 {
-					t.Logf("stderr: %s", stderrBuf.String())
-				}
-				require.NoError(t, err, "command should succeed: %s", tc.command)
-				assert.Equal(t, tc.expect, string(output), "output mismatch for: %s", tc.command)
-			case <-time.After(5 * time.Second):
-				t.Fatalf("command timed out: %s", tc.command)
-			}
-		})
-	}
-}
-
-// setupProxySSHClient creates a full proxy test environment and returns
-// an SSH client connected through the proxy to a backend NetBird SSH server.
-func setupProxySSHClient(t *testing.T) (*cryptossh.Client, func()) {
-	t.Helper()
-
-	const (
-		issuer   = "https://test-issuer.example.com"
-		audience = "test-audience"
-	)
-
-	jwksServer, privateKey, jwksURL := setupJWKSServer(t)
-
-	hostKey, err := nbssh.GeneratePrivateKey(nbssh.ED25519)
-	require.NoError(t, err)
-	hostPubKey, err := nbssh.GeneratePublicKey(hostKey)
-	require.NoError(t, err)
-
-	serverConfig := &server.Config{
-		HostKeyPEM: hostKey,
-		JWT: &server.JWTConfig{
-			Issuer:       issuer,
-			Audiences:    []string{audience},
-			KeysLocation: jwksURL,
-		},
-	}
-	sshServer := server.New(serverConfig)
-	sshServer.SetAllowRootLogin(true)
-
-	testUsername := testutil.GetTestUsername(t)
-	testJWTUser := "test-username"
-	testUserHash, err := sshuserhash.HashUserID(testJWTUser)
-	require.NoError(t, err)
-
-	authConfig := &sshauth.Config{
-		UserIDClaim:     sshauth.DefaultUserIDClaim,
-		AuthorizedUsers: []sshuserhash.UserIDHash{testUserHash},
-		MachineUsers: map[string][]uint32{
-			testUsername: {0},
-		},
-	}
-	sshServer.UpdateSSHAuth(authConfig)
-
-	sshServerAddr := server.StartTestServer(t, sshServer)
-
-	mockDaemon := startMockDaemon(t)
-
-	host, portStr, err := net.SplitHostPort(sshServerAddr)
-	require.NoError(t, err)
-	port, err := strconv.Atoi(portStr)
-	require.NoError(t, err)
-
-	mockDaemon.setHostKey(host, hostPubKey)
-
-	validToken := generateValidJWT(t, privateKey, issuer, audience, testJWTUser)
-	mockDaemon.setJWTToken(validToken)
-
-	proxyInstance, err := New(mockDaemon.addr, host, port, io.Discard, nil)
-	require.NoError(t, err)
-
-	origStdin := os.Stdin
-	origStdout := os.Stdout
-
-	stdinReader, stdinWriter, err := os.Pipe()
-	require.NoError(t, err)
-	stdoutReader, stdoutWriter, err := os.Pipe()
-	require.NoError(t, err)
-
-	os.Stdin = stdinReader
-	os.Stdout = stdoutWriter
-
-	clientConn, proxyConn := net.Pipe()
-
-	go func() { _, _ = io.Copy(stdinWriter, proxyConn) }()
-	go func() { _, _ = io.Copy(proxyConn, stdoutReader) }()
-
-	ctx, cancel := context.WithTimeout(context.Background(), 10*time.Second)
-
-	go func() {
-		_ = proxyInstance.Connect(ctx)
-	}()
-
-	sshConfig := &cryptossh.ClientConfig{
-		User:            testutil.GetTestUsername(t),
-		Auth:            []cryptossh.AuthMethod{},
-		HostKeyCallback: cryptossh.InsecureIgnoreHostKey(),
-		Timeout:         5 * time.Second,
-	}
-
-	sshClientConn, chans, reqs, err := cryptossh.NewClientConn(clientConn, "test", sshConfig)
-	require.NoError(t, err)
-
-	client := cryptossh.NewClient(sshClientConn, chans, reqs)
-
-	cleanupFn := func() {
-		_ = client.Close()
-		_ = clientConn.Close()
-		cancel()
-		os.Stdin = origStdin
-		os.Stdout = origStdout
-		_ = sshServer.Stop()
-		mockDaemon.stop()
-		jwksServer.Close()
-	}
-
-	return client, cleanupFn
-}
-
 type mockDaemonServer struct {
 	proto.UnimplementedDaemonServiceServer
 	hostKeys map[string][]byte
@@ -492,10 +150,6 @@ func (m *mockDaemon) setHostKey(addr string, pubKey []byte) {
 	m.impl.hostKeys[addr] = pubKey
 }
 
-func (m *mockDaemon) setJWTToken(token string) {
-	m.impl.jwtToken = token
-}
-
 func (m *mockDaemon) stop() {
 	if m.server != nil {
 		m.server.Stop()
@@ -508,63 +162,3 @@ func mustParsePublicKey(t *testing.T, pubKeyBytes []byte) cryptossh.PublicKey {
 	require.NoError(t, err)
 	return pubKey
 }
-
-func setupJWKSServer(t *testing.T) (*httptest.Server, *rsa.PrivateKey, string) {
-	t.Helper()
-	privateKey, jwksJSON := generateTestJWKS(t)
-
-	server := httptest.NewServer(http.HandlerFunc(func(w http.ResponseWriter, r *http.Request) {
-		w.Header().Set("Content-Type", "application/json")
-		if _, err := w.Write(jwksJSON); err != nil {
-			http.Error(w, err.Error(), http.StatusInternalServerError)
-		}
-	}))
-
-	return server, privateKey, server.URL
-}
-
-func generateTestJWKS(t *testing.T) (*rsa.PrivateKey, []byte) {
-	t.Helper()
-	privateKey, err := rsa.GenerateKey(rand.Reader, 2048)
-	require.NoError(t, err)
-
-	publicKey := &privateKey.PublicKey
-	n := publicKey.N.Bytes()
-	e := publicKey.E
-
-	jwk := nbjwt.JSONWebKey{
-		Kty: "RSA",
-		Kid: "test-key-id",
-		Use: "sig",
-		N:   base64.RawURLEncoding.EncodeToString(n),
-		E:   base64.RawURLEncoding.EncodeToString(big.NewInt(int64(e)).Bytes()),
-	}
-
-	jwks := nbjwt.Jwks{
-		Keys: []nbjwt.JSONWebKey{jwk},
-	}
-
-	jwksJSON, err := json.Marshal(jwks)
-	require.NoError(t, err)
-
-	return privateKey, jwksJSON
-}
-
-func generateValidJWT(t *testing.T, privateKey *rsa.PrivateKey, issuer, audience string, user string) string {
-	t.Helper()
-	claims := jwt.MapClaims{
-		"iss": issuer,
-		"aud": audience,
-		"sub": user,
-		"exp": time.Now().Add(time.Hour).Unix(),
-		"iat": time.Now().Unix(),
-	}
-
-	token := jwt.NewWithClaims(jwt.SigningMethodRS256, claims)
-	token.Header["kid"] = "test-key-id"
-
-	tokenString, err := token.SignedString(privateKey)
-	require.NoError(t, err)
-
-	return tokenString
-}

client/ssh/server/executor_unix_privileged_test.go

@@ -0,0 +1,66 @@
+//go:build unix && privileged
+
+package server
+
+import (
+	"context"
+	"testing"
+
+	"github.com/stretchr/testify/assert"
+	"github.com/stretchr/testify/require"
+)
+
+func TestPrivilegeDropper_CreateExecutorCommand(t *testing.T) {
+	pd := NewPrivilegeDropper()
+
+	config := ExecutorConfig{
+		UID:        1000,
+		GID:        1000,
+		Groups:     []uint32{1000, 1001},
+		WorkingDir: "/home/testuser",
+		Shell:      "/bin/bash",
+		Command:    "ls -la",
+	}
+
+	cmd, err := pd.CreateExecutorCommand(context.Background(), config)
+	require.NoError(t, err)
+	require.NotNil(t, cmd)
+
+	// Verify the command is calling netbird ssh exec
+	assert.Contains(t, cmd.Args, "ssh")
+	assert.Contains(t, cmd.Args, "exec")
+	assert.Contains(t, cmd.Args, "--uid")
+	assert.Contains(t, cmd.Args, "1000")
+	assert.Contains(t, cmd.Args, "--gid")
+	assert.Contains(t, cmd.Args, "1000")
+	assert.Contains(t, cmd.Args, "--groups")
+	assert.Contains(t, cmd.Args, "1000")
+	assert.Contains(t, cmd.Args, "1001")
+	assert.Contains(t, cmd.Args, "--working-dir")
+	assert.Contains(t, cmd.Args, "/home/testuser")
+	assert.Contains(t, cmd.Args, "--shell")
+	assert.Contains(t, cmd.Args, "/bin/bash")
+	assert.Contains(t, cmd.Args, "--cmd")
+	assert.Contains(t, cmd.Args, "ls -la")
+}
+
+func TestPrivilegeDropper_CreateExecutorCommandInteractive(t *testing.T) {
+	pd := NewPrivilegeDropper()
+
+	config := ExecutorConfig{
+		UID:        1000,
+		GID:        1000,
+		Groups:     []uint32{1000},
+		WorkingDir: "/home/testuser",
+		Shell:      "/bin/bash",
+		Command:    "",
+	}
+
+	cmd, err := pd.CreateExecutorCommand(context.Background(), config)
+	require.NoError(t, err)
+	require.NotNil(t, cmd)
+
+	// Verify no command mode (command is empty so no --cmd flag)
+	assert.NotContains(t, cmd.Args, "--cmd")
+	assert.NotContains(t, cmd.Args, "--interactive")
+}

client/ssh/server/executor_unix_test.go

@@ -73,61 +73,6 @@ func TestPrivilegeDropper_ValidatePrivileges(t *testing.T) {
 	}
 }
 
-func TestPrivilegeDropper_CreateExecutorCommand(t *testing.T) {
-	pd := NewPrivilegeDropper()
-
-	config := ExecutorConfig{
-		UID:        1000,
-		GID:        1000,
-		Groups:     []uint32{1000, 1001},
-		WorkingDir: "/home/testuser",
-		Shell:      "/bin/bash",
-		Command:    "ls -la",
-	}
-
-	cmd, err := pd.CreateExecutorCommand(context.Background(), config)
-	require.NoError(t, err)
-	require.NotNil(t, cmd)
-
-	// Verify the command is calling netbird ssh exec
-	assert.Contains(t, cmd.Args, "ssh")
-	assert.Contains(t, cmd.Args, "exec")
-	assert.Contains(t, cmd.Args, "--uid")
-	assert.Contains(t, cmd.Args, "1000")
-	assert.Contains(t, cmd.Args, "--gid")
-	assert.Contains(t, cmd.Args, "1000")
-	assert.Contains(t, cmd.Args, "--groups")
-	assert.Contains(t, cmd.Args, "1000")
-	assert.Contains(t, cmd.Args, "1001")
-	assert.Contains(t, cmd.Args, "--working-dir")
-	assert.Contains(t, cmd.Args, "/home/testuser")
-	assert.Contains(t, cmd.Args, "--shell")
-	assert.Contains(t, cmd.Args, "/bin/bash")
-	assert.Contains(t, cmd.Args, "--cmd")
-	assert.Contains(t, cmd.Args, "ls -la")
-}
-
-func TestPrivilegeDropper_CreateExecutorCommandInteractive(t *testing.T) {
-	pd := NewPrivilegeDropper()
-
-	config := ExecutorConfig{
-		UID:        1000,
-		GID:        1000,
-		Groups:     []uint32{1000},
-		WorkingDir: "/home/testuser",
-		Shell:      "/bin/bash",
-		Command:    "",
-	}
-
-	cmd, err := pd.CreateExecutorCommand(context.Background(), config)
-	require.NoError(t, err)
-	require.NotNil(t, cmd)
-
-	// Verify no command mode (command is empty so no --cmd flag)
-	assert.NotContains(t, cmd.Args, "--cmd")
-	assert.NotContains(t, cmd.Args, "--interactive")
-}
-
 // TestPrivilegeDropper_ActualPrivilegeDrop tests actual privilege dropping
 // This test requires root privileges and will be skipped if not running as root
 func TestPrivilegeDropper_ActualPrivilegeDrop(t *testing.T) {

client/system/info.go

@@ -3,6 +3,7 @@ package system
 import (
 	"context"
 	"net/netip"
+	"slices"
 	"strings"
 
 	log "github.com/sirupsen/logrus"
@@ -121,6 +122,23 @@ func (i *Info) SetFlags(
 	}
 }
 
+// removeAddresses drops network addresses whose IP matches any of the given
+// addresses, regardless of prefix length. Used to exclude the NetBird overlay
+// address, which otherwise churns the meta as the interface comes and goes.
+func (i *Info) removeAddresses(ips ...netip.Addr) {
+	if len(ips) == 0 {
+		return
+	}
+	filtered := i.NetworkAddresses[:0]
+	for _, addr := range i.NetworkAddresses {
+		if slices.Contains(ips, addr.NetIP.Addr()) {
+			continue
+		}
+		filtered = append(filtered, addr)
+	}
+	i.NetworkAddresses = filtered
+}
+
 // extractUserAgent extracts Netbird's agent (client) name and version from the outgoing context
 func extractUserAgent(ctx context.Context) string {
 	md, hasMeta := metadata.FromOutgoingContext(ctx)
@@ -147,7 +165,9 @@ func extractDeviceName(ctx context.Context, defaultName string) string {
 }
 
 // GetInfoWithChecks retrieves and parses the system information with applied checks.
-func GetInfoWithChecks(ctx context.Context, checks []*proto.Checks) (*Info, error) {
+// excludeIPs are dropped from the reported network addresses (e.g. our own
+// WireGuard overlay address, which otherwise churns the peer meta).
+func GetInfoWithChecks(ctx context.Context, checks []*proto.Checks, excludeIPs ...netip.Addr) (*Info, error) {
 	log.Debugf("gathering system information with checks: %d", len(checks))
 	processCheckPaths := make([]string, 0)
 	for _, check := range checks {
@@ -162,6 +182,7 @@ func GetInfoWithChecks(ctx context.Context, checks []*proto.Checks) (*Info, erro
 
 	info := GetInfo(ctx)
 	info.Files = files
+	info.removeAddresses(excludeIPs...)
 
 	log.Debugf("all system information gathered successfully")
 	return info, nil

client/system/info_test.go

@@ -2,6 +2,7 @@ package system
 
 import (
 	"context"
+	"net/netip"
 	"testing"
 
 	"github.com/stretchr/testify/assert"
@@ -43,3 +44,42 @@ func Test_NetAddresses(t *testing.T) {
 		t.Errorf("no network addresses found")
 	}
 }
+
+func TestInfo_RemoveAddresses(t *testing.T) {
+	addr := func(cidr string) NetworkAddress {
+		return NetworkAddress{NetIP: netip.MustParsePrefix(cidr)}
+	}
+
+	info := &Info{
+		NetworkAddresses: []NetworkAddress{
+			addr("192.168.1.7/24"),
+			addr("100.76.70.97/32"),                          // overlay v4 (host mask /32)
+			addr("2001:818:c51b:4800:845:a65d:ae6f:623f/64"), // real global v6
+			addr("fd00:1234::1/64"),                          // overlay v6
+		},
+	}
+
+	// Overlay addresses as the engine knows them, with a different mask (/16, /64).
+	info.removeAddresses(
+		netip.MustParseAddr("100.76.70.97"),
+		netip.MustParseAddr("fd00:1234::1"),
+	)
+
+	want := []string{"192.168.1.7/24", "2001:818:c51b:4800:845:a65d:ae6f:623f/64"}
+	if len(info.NetworkAddresses) != len(want) {
+		t.Fatalf("got %d addresses, want %d: %v", len(info.NetworkAddresses), len(want), info.NetworkAddresses)
+	}
+	for i, w := range want {
+		if got := info.NetworkAddresses[i].NetIP.String(); got != w {
+			t.Errorf("address[%d] = %s, want %s", i, got, w)
+		}
+	}
+}
+
+func TestInfo_RemoveAddresses_NoOp(t *testing.T) {
+	info := &Info{NetworkAddresses: []NetworkAddress{{NetIP: netip.MustParsePrefix("10.0.0.1/24")}}}
+	info.removeAddresses()
+	if len(info.NetworkAddresses) != 1 {
+		t.Errorf("expected no change with empty input, got %v", info.NetworkAddresses)
+	}
+}

client/system/network_addr.go

@@ -46,7 +46,9 @@ func toNetworkAddress(address net.Addr, mac string) (NetworkAddress, bool) {
 	if !ok {
 		return NetworkAddress{}, false
 	}
-	if ipNet.IP.IsLoopback() {
+	// Skip link-local and multicast: they carry no routable peer info and the
+	// IPv6 link-local of a flapping NIC churns the meta on every up/down.
+	if ipNet.IP.IsLoopback() || ipNet.IP.IsLinkLocalUnicast() || ipNet.IP.IsMulticast() {
 		return NetworkAddress{}, false
 	}
 	prefix, err := netip.ParsePrefix(ipNet.String())

client/system/network_addr_test.go

@@ -0,0 +1,45 @@
+//go:build !ios
+
+package system
+
+import (
+	"net"
+	"testing"
+)
+
+func mustIPNet(t *testing.T, cidr string) *net.IPNet {
+	t.Helper()
+	ip, ipNet, err := net.ParseCIDR(cidr)
+	if err != nil {
+		t.Fatalf("parse %q: %v", cidr, err)
+	}
+	ipNet.IP = ip
+	return ipNet
+}
+
+func TestToNetworkAddress_Filtering(t *testing.T) {
+	const mac = "c8:4b:d6:b6:04:ac"
+
+	tests := []struct {
+		name string
+		cidr string
+		want bool
+	}{
+		{"ipv4 global", "10.65.16.181/23", true},
+		{"ipv6 global", "2620:52:0:4110:102d:6a98:ee75:8b92/64", true},
+		{"ipv4 loopback", "127.0.0.1/8", false},
+		{"ipv6 loopback", "::1/128", false},
+		{"ipv6 link-local", "fe80::871:4c25:23d7:2529/64", false},
+		{"ipv4 link-local", "169.254.1.2/16", false},
+		{"ipv6 multicast", "ff02::1/128", false},
+	}
+
+	for _, tt := range tests {
+		t.Run(tt.name, func(t *testing.T) {
+			_, got := toNetworkAddress(mustIPNet(t, tt.cidr), mac)
+			if got != tt.want {
+				t.Errorf("toNetworkAddress(%s) ok = %v, want %v", tt.cidr, got, tt.want)
+			}
+		})
+	}
+}

client/testutil/privileged/runner_test.go

@@ -0,0 +1,196 @@
+//go:build privileged && (linux || darwin)
+
+// Package privileged provides a self-hosting harness that runs the repo's
+// privileged-tagged test suite inside a --privileged --cap-add=NET_ADMIN
+// container, so developers can exercise the root/system-mutating tests on a
+// non-root host with a single `go test` invocation.
+package privileged
+
+import (
+	"bytes"
+	"context"
+	"fmt"
+	"os"
+	"os/exec"
+	"path/filepath"
+	"strings"
+	"testing"
+	"time"
+
+	"github.com/moby/moby/api/types/container"
+	"github.com/ory/dockertest/v4"
+)
+
+// containerImage / containerTag match the image used by the CI privileged job
+// (.github/workflows/golang-test-linux.yml, test_client_on_docker).
+const (
+	containerImage = "golang"
+	containerTag   = "1.25-alpine"
+)
+
+const (
+	containerWorkdir    = "/app"
+	containerGoCache    = "/root/.cache/go-build"
+	containerGoModCache = "/go/pkg/mod"
+)
+
+// alpinePackages are the build/runtime deps the privileged tests need, mirroring
+// the CI container setup.
+const alpinePackages = "ca-certificates iptables ip6tables dbus dbus-dev libpcap-dev build-base"
+
+// privilegedTestPackages is the package list the suite runs, excluding the
+// server-side trees and UI/upload helpers, matching the CI Docker job's filter.
+const privilegedTestPackages = `go list -buildvcs=false ./... | grep -v -e /management -e /signal -e /relay -e /proxy -e /combined -e /client/ui -e /upload-server`
+
+// testWriter forwards container output to the test log line by line.
+type testWriter struct{ t *testing.T }
+
+func (w testWriter) Write(p []byte) (int, error) {
+	for _, line := range strings.Split(strings.TrimRight(string(p), "\n"), "\n") {
+		w.t.Log(line)
+	}
+	return len(p), nil
+}
+
+// TestRunPrivilegedSuiteInDocker spins up a privileged container, mounts the repo,
+// and runs `go test -tags 'devcert privileged'` inside it. When already running
+// inside that container (DOCKER_CI=true) it returns immediately so the real
+// privileged tests in the suite execute in place instead of recursing.
+func TestRunPrivilegedSuiteInDocker(t *testing.T) {
+	if os.Getenv("DOCKER_CI") == "true" {
+		t.Skip("inside privileged container, skipping container spawn; privileged tests run in place")
+	}
+
+	repoRoot, err := findRepoRoot()
+	if err != nil {
+		t.Fatalf("locate repo root: %v", err)
+	}
+	goCache, goModCache := hostGoCaches(t)
+
+	// dockertest reads DOCKER_HOST; point it at the active context's socket when
+	// the default one is absent (macOS Docker Desktop, Colima, OrbStack).
+	if host := dockerHost(); host != "" {
+		t.Setenv("DOCKER_HOST", host)
+	}
+
+	// NewPoolT registers container cleanup via t.Cleanup automatically.
+	pool := dockertest.NewPoolT(t, "", dockertest.WithMaxWait(30*time.Minute))
+
+	// Keep the container alive so the suite runs via Exec, which yields a clean
+	// exit code (the v4 Resource API exposes no container wait/exit-code).
+	resource := pool.RunT(t, containerImage,
+		dockertest.WithTag(containerTag),
+		dockertest.WithWorkingDir(containerWorkdir),
+		dockertest.WithMounts([]string{
+			repoRoot + ":" + containerWorkdir,
+			goCache + ":" + containerGoCache,
+			goModCache + ":" + containerGoModCache,
+		}),
+		dockertest.WithEnv([]string{
+			"CGO_ENABLED=1",
+			"CI=true",
+			"DOCKER_CI=true",
+			"CONTAINER=true",
+			"GOCACHE=" + containerGoCache,
+			"GOMODCACHE=" + containerGoModCache,
+		}),
+		dockertest.WithCmd([]string{"sleep", "infinity"}),
+		dockertest.WithHostConfig(func(hc *container.HostConfig) {
+			hc.Privileged = true
+			hc.CapAdd = []string{"NET_ADMIN"}
+		}),
+		dockertest.WithoutReuse(),
+	)
+
+	ctx, cancel := context.WithTimeout(context.Background(), 30*time.Minute)
+	defer cancel()
+
+	result, err := resource.Exec(ctx, []string{"sh", "-c", buildTestScript()})
+	if err != nil {
+		t.Fatalf("run privileged suite in container: %v", err)
+	}
+
+	w := testWriter{t}
+	_, _ = w.Write([]byte(result.StdOut))
+	_, _ = w.Write([]byte(result.StdErr))
+
+	if result.ExitCode != 0 {
+		t.Fatalf("privileged test suite failed in container (exit code %d)", result.ExitCode)
+	}
+}
+
+// findRepoRoot walks up from the test's working directory to the module root.
+func findRepoRoot() (string, error) {
+	dir, err := os.Getwd()
+	if err != nil {
+		return "", err
+	}
+	for {
+		if _, statErr := os.Stat(filepath.Join(dir, "go.mod")); statErr == nil {
+			return dir, nil
+		}
+		parent := filepath.Dir(dir)
+		if parent == dir {
+			return "", fmt.Errorf("go.mod not found above %s", dir)
+		}
+		dir = parent
+	}
+}
+
+// dockerHost returns a DOCKER_HOST override when the default socket is missing.
+// An empty result means the caller should leave DOCKER_HOST untouched (it is
+// already set, or the default unix socket exists). When neither is present
+// (common on macOS Docker Desktop, Colima and OrbStack, which use a per-user
+// socket), it resolves the active docker context's endpoint.
+func dockerHost() string {
+	if os.Getenv("DOCKER_HOST") != "" {
+		return ""
+	}
+	if _, err := os.Stat("/var/run/docker.sock"); err == nil {
+		return ""
+	}
+
+	out, err := exec.Command("docker", "context", "inspect", "-f", "{{.Endpoints.docker.Host}}").Output()
+	if err != nil {
+		return ""
+	}
+	return strings.TrimSpace(string(out))
+}
+
+// hostGoCaches resolves the host GOCACHE/GOMODCACHE so the container reuses the
+// existing build/module cache for speed.
+func hostGoCaches(t *testing.T) (string, string) {
+	t.Helper()
+	return goEnv(t, "GOCACHE"), goEnv(t, "GOMODCACHE")
+}
+
+func goEnv(t *testing.T, key string) string {
+	t.Helper()
+	var out bytes.Buffer
+	cmd := exec.Command("go", "env", key)
+	cmd.Stdout = &out
+	if err := cmd.Run(); err != nil {
+		t.Fatalf("go env %s: %v", key, err)
+	}
+	return strings.TrimSpace(out.String())
+}
+
+// buildTestScript builds the in-container command. PRIV_PKGS overrides the package
+// list (default: the full filtered set); PRIV_RUN adds a -run test-name filter.
+// Both empty reproduces the full privileged suite.
+func buildTestScript() string {
+	pkgs := privilegedTestPackages + " | xargs"
+	if p := os.Getenv("PRIV_PKGS"); p != "" {
+		pkgs = "echo " + p + " | xargs"
+	}
+
+	runFilter := ""
+	if r := os.Getenv("PRIV_RUN"); r != "" {
+		runFilter = "-run '" + r + "' "
+	}
+
+	return fmt.Sprintf(
+		"apk update >/dev/null && apk add --no-cache %s >/dev/null && %s go test -buildvcs=false -tags 'devcert privileged' %s-v -timeout 20m -p 1",
+		alpinePackages, pkgs, runFilter,
+	)
+}

client/ui/client_ui.go

@@ -418,7 +418,14 @@ func newServiceClient(args *newServiceClientArgs) *serviceClient {
 	case args.showProfiles:
 		s.showProfilesUI()
 	case args.showQuickActions:
-		s.showQuickActionsUI()
+		// Suppress the on-boot Quick Actions popup when the daemon
+		// reports DisableAutoConnect=true — that flag carries both the
+		// user's "Connect on Startup = off" preference AND any MDM-
+		// enforced override (applyMDMPolicy writes the policy value
+		// into the same Config field). See netbirdio/netbird#5744.
+		if !s.disableAutoConnectFromDaemon() {
+			s.showQuickActionsUI()
+		}
 	case args.showUpdate:
 		s.showUpdateProgress(ctx, args.showUpdateVersion)
 	}
@@ -1338,6 +1345,40 @@ func (s *serviceClient) getFeatures() (*proto.GetFeaturesResponse, error) {
 	return features, nil
 }
 
+// disableAutoConnectFromDaemon returns true when the daemon reports
+// the active profile has DisableAutoConnect=true. Used by the
+// --quick-actions startup path to suppress the on-boot popup when the
+// user (or an MDM admin) opted out of auto-connecting; both cases
+// converge on the same Config field because applyMDMPolicy writes the
+// policy value into it. Returns false on any RPC / lookup failure so a
+// daemon hiccup does not silently swallow the popup.
+func (s *serviceClient) disableAutoConnectFromDaemon() bool {
+	activeProf, err := s.profileManager.GetActiveProfile()
+	if err != nil {
+		log.Warnf("disableAutoConnectFromDaemon: get active profile: %v", err)
+		return false
+	}
+	currUser, err := user.Current()
+	if err != nil {
+		log.Warnf("disableAutoConnectFromDaemon: get current user: %v", err)
+		return false
+	}
+	conn, err := s.getSrvClient(failFastTimeout)
+	if err != nil {
+		log.Warnf("disableAutoConnectFromDaemon: get daemon client: %v", err)
+		return false
+	}
+	srvCfg, err := conn.GetConfig(s.ctx, &proto.GetConfigRequest{
+		ProfileName: activeProf.ID.String(),
+		Username:    currUser.Username,
+	})
+	if err != nil {
+		log.Warnf("disableAutoConnectFromDaemon: GetConfig RPC: %v", err)
+		return false
+	}
+	return srvCfg.GetDisableAutoConnect()
+}
+
 // getSrvConfig from the service to show it in the settings window.
 func (s *serviceClient) getSrvConfig() {
 	s.managementURL = profilemanager.DefaultManagementURL

docs/testing-privileged.md

@@ -0,0 +1,78 @@
+# Privileged tests
+
+Some tests in this repo need `root` or mutate host network state: they create
+TUN/WireGuard interfaces, open netlink/raw sockets, run eBPF programs, or shell
+out to `ip`/`iptables`/`nft`/`ifconfig`/`route`. Running them on a developer
+machine would require `sudo` and could leave stray interfaces or routes behind.
+
+These tests are gated behind the **`privileged` build tag** so the default test
+run is host-safe.
+
+## Running tests
+
+```bash
+# Host-safe: excludes privileged tests. Runs as a normal user, no sudo.
+make test-unit
+# equivalently:
+go test -tags devcert ./...
+
+# Privileged suite: runs the privileged-tagged tests inside a
+# --privileged --cap-add=NET_ADMIN container (requires Docker).
+make test-privileged
+
+# Narrow the container run to a single test / package:
+PRIV_RUN=TestNftablesManager PRIV_PKGS=./client/firewall/nftables/... make test-privileged
+```
+
+`PRIV_RUN` adds a `-run` test-name filter and `PRIV_PKGS` overrides the package
+list; both are optional and default to the full privileged suite.
+
+`make test-privileged` invokes the `ory/dockertest` harness in
+`client/testutil/privileged/`. The harness:
+
+1. Skips immediately when it detects it is already inside the container
+   (`DOCKER_CI=true`), so the privileged tests run in place instead of recursing.
+2. Otherwise spins up a `golang:1.25-alpine` container (matching CI),
+   bind-mounts the repo and the host Go build/module caches, installs the
+   required packages, and runs `go test -tags 'devcert privileged'` over the
+   client packages.
+3. Streams the container's output to the test log and fails if the suite fails.
+
+## Adding a privileged test
+
+A test is privileged if it does any of:
+
+- creates a real interface via `iface.NewWGIFace(...).Create()`,
+- opens a netlink or raw socket that hard-fails without `CAP_NET_ADMIN`,
+- runs an eBPF program (`ebpf.*.Listen()`),
+- shells out to `ip`, `iptables`, `nft`, `ifconfig`, or `route` to change state.
+
+Add the tag to the **top** of the file, combined with any existing platform
+constraint:
+
+```go
+//go:build privileged && linux
+
+package foo
+```
+
+If a file mixes privileged and pure-logic tests, **split it**: keep the pure
+tests (and any shared data — type/var declarations, table-driven `testCases`,
+helper interfaces) in an untagged file, and move the privileged tests into a
+`*_privileged_test.go` file with the tag. Shared declarations must stay untagged,
+otherwise the unprivileged files in the package will not compile.
+
+Always verify both build modes compile on every target platform:
+
+```bash
+go vet -tags devcert ./...
+go vet -tags 'devcert privileged' ./...
+```
+
+## CI
+
+- The `Client / Unit` job runs `go test -tags devcert` with **no** `sudo` — only
+  host-safe tests.
+- The `Client (Docker) / Unit` job runs `go test -tags 'devcert privileged'`
+  inside a `--privileged --cap-add=NET_ADMIN` container, which is where the
+  privileged tests actually execute.

go.mod

@@ -78,10 +78,12 @@ require (
 	github.com/mdp/qrterminal/v3 v3.2.1
 	github.com/miekg/dns v1.1.72
 	github.com/mitchellh/hashstructure/v2 v2.0.2
+	github.com/moby/moby/api v1.54.1
 	github.com/netbirdio/management-integrations/integrations v0.0.0-20260416123949-2355d972be42
 	github.com/netbirdio/signal-dispatcher/dispatcher v0.0.0-20250805121659-6b4ac470ca45
 	github.com/oapi-codegen/runtime v1.1.2
 	github.com/okta/okta-sdk-golang/v2 v2.18.0
+	github.com/ory/dockertest/v4 v4.0.0
 	github.com/oschwald/maxminddb-golang v1.12.0
 	github.com/patrickmn/go-cache v2.1.0+incompatible
 	github.com/petermattis/goid v0.0.0-20250303134427-723919f7f203
@@ -145,7 +147,7 @@ require (
 	dario.cat/mergo v1.0.1 // indirect
 	filippo.io/edwards25519 v1.1.1 // indirect
 	github.com/AppsFlyer/go-sundheit v0.6.0 // indirect
-	github.com/Azure/go-ansiterm v0.0.0-20230124172434-306776ec8161 // indirect
+	github.com/Azure/go-ansiterm v0.0.0-20250102033503-faa5f7b0171c // indirect
 	github.com/Azure/go-ntlmssp v0.1.0 // indirect
 	github.com/BurntSushi/toml v1.5.0 // indirect
 	github.com/Masterminds/goutils v1.1.1 // indirect
@@ -177,6 +179,8 @@ require (
 	github.com/caddyserver/zerossl v0.1.3 // indirect
 	github.com/cenkalti/backoff/v5 v5.0.3 // 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/containerd/platforms v0.2.1 // indirect
 	github.com/cpuguy83/dockercfg v0.3.2 // indirect
@@ -271,11 +275,12 @@ require (
 	github.com/mitchellh/mapstructure v1.5.0 // indirect
 	github.com/mitchellh/reflectwalk v1.0.2 // indirect
 	github.com/moby/docker-image-spec v1.3.1 // indirect
+	github.com/moby/moby/client v0.4.0 // indirect
 	github.com/moby/patternmatcher v0.6.0 // indirect
 	github.com/moby/sys/sequential v0.5.0 // indirect
 	github.com/moby/sys/user v0.3.0 // indirect
 	github.com/moby/sys/userns v0.1.0 // indirect
-	github.com/moby/term v0.5.0 // indirect
+	github.com/moby/term v0.5.2 // indirect
 	github.com/morikuni/aec v1.0.0 // indirect
 	github.com/munnerz/goautoneg v0.0.0-20191010083416-a7dc8b61c822 // indirect
 	github.com/nfnt/resize v0.0.0-20180221191011-83c6a9932646 // indirect
@@ -341,7 +346,7 @@ replace github.com/kardianos/service => github.com/netbirdio/service v0.0.0-2024
 
 replace github.com/getlantern/systray => github.com/netbirdio/systray v0.0.0-20231030152038-ef1ed2a27949
 
-replace golang.zx2c4.com/wireguard => github.com/netbirdio/wireguard-go v0.0.0-20260523085312-4b4a4e36017f
+replace golang.zx2c4.com/wireguard => github.com/netbirdio/wireguard-go v0.0.0-20260628102922-2834bebf6c1a
 
 replace github.com/cloudflare/circl => codeberg.org/cunicu/circl v0.0.0-20230801113412-fec58fc7b5f6
 

go.sum

@@ -23,8 +23,8 @@ github.com/AdaLogics/go-fuzz-headers v0.0.0-20230811130428-ced1acdcaa24 h1:bvDV9
 github.com/AdaLogics/go-fuzz-headers v0.0.0-20230811130428-ced1acdcaa24/go.mod h1:8o94RPi1/7XTJvwPpRSzSUedZrtlirdB3r9Z20bi2f8=
 github.com/AppsFlyer/go-sundheit v0.6.0 h1:d2hBvCjBSb2lUsEWGfPigr4MCOt04sxB+Rppl0yUMSk=
 github.com/AppsFlyer/go-sundheit v0.6.0/go.mod h1:LDdBHD6tQBtmHsdW+i1GwdTt6Wqc0qazf5ZEJVTbTME=
-github.com/Azure/go-ansiterm v0.0.0-20230124172434-306776ec8161 h1:L/gRVlceqvL25UVaW/CKtUDjefjrs0SPonmDGUVOYP0=
+github.com/Azure/go-ansiterm v0.0.0-20250102033503-faa5f7b0171c h1:udKWzYgxTojEKWjV8V+WSxDXJ4NFATAsZjh8iIbsQIg=
-github.com/Azure/go-ansiterm v0.0.0-20230124172434-306776ec8161/go.mod h1:xomTg63KZ2rFqZQzSB4Vz2SUXa1BpHTVz9L5PTmPC4E=
+github.com/Azure/go-ansiterm v0.0.0-20250102033503-faa5f7b0171c/go.mod h1:xomTg63KZ2rFqZQzSB4Vz2SUXa1BpHTVz9L5PTmPC4E=
 github.com/Azure/go-ntlmssp v0.1.0 h1:DjFo6YtWzNqNvQdrwEyr/e4nhU3vRiwenz5QX7sFz+A=
 github.com/Azure/go-ntlmssp v0.1.0/go.mod h1:NYqdhxd/8aAct/s4qSYZEerdPuH1liG2/X9DiVTbhpk=
 github.com/BurntSushi/toml v1.5.0 h1:W5quZX/G/csjUnuI8SUYlsHs9M38FC7znL0lIO+DvMg=
@@ -117,6 +117,10 @@ github.com/cilium/ebpf v0.19.0 h1:Ro/rE64RmFBeA9FGjcTc+KmCeY6jXmryu6FfnzPRIao=
 github.com/cilium/ebpf v0.19.0/go.mod h1:fLCgMo3l8tZmAdM3B2XqdFzXBpwkcSTroaVqN08OWVY=
 github.com/coder/websocket v1.8.14 h1:9L0p0iKiNOibykf283eHkKUHHrpG7f65OE3BhhO7v9g=
 github.com/coder/websocket v1.8.14/go.mod h1:NX3SzP+inril6yawo5CQXx8+fk145lPDC6pumgx0mVg=
+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/containerd/platforms v0.2.1 h1:zvwtM3rz2YHPQsF2CHYM8+KtB5dvhISiXh5ZpSBQv6A=
@@ -480,6 +484,10 @@ github.com/mitchellh/reflectwalk v1.0.2 h1:G2LzWKi524PWgd3mLHV8Y5k7s6XUvT0Gef6zx
 github.com/mitchellh/reflectwalk v1.0.2/go.mod h1:mSTlrgnPZtwu0c4WaC2kGObEpuNDbx0jmZXqmk4esnw=
 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/moby/api v1.54.1 h1:TqVzuJkOLsgLDDwNLmYqACUuTehOHRGKiPhvH8V3Nn4=
+github.com/moby/moby/api v1.54.1/go.mod h1:+RQ6wluLwtYaTd1WnPLykIDPekkuyD/ROWQClE83pzs=
+github.com/moby/moby/client v0.4.0 h1:S+2XegzHQrrvTCvF6s5HFzcrywWQmuVnhOXe2kiWjIw=
+github.com/moby/moby/client v0.4.0/go.mod h1:QWPbvWchQbxBNdaLSpoKpCdf5E+WxFAgNHogCWDoa7g=
 github.com/moby/patternmatcher v0.6.0 h1:GmP9lR19aU5GqSSFko+5pRqHi+Ohk1O69aFiKkVGiPk=
 github.com/moby/patternmatcher v0.6.0/go.mod h1:hDPoyOpDY7OrrMDLaYoY3hf52gNCR/YOUYxkhApJIxc=
 github.com/moby/sys/sequential v0.5.0 h1:OPvI35Lzn9K04PBbCLW0g4LcFAJgHsvXsRyewg5lXtc=
@@ -488,8 +496,8 @@ github.com/moby/sys/user v0.3.0 h1:9ni5DlcW5an3SvRSx4MouotOygvzaXbaSrc/wGDFWPo=
 github.com/moby/sys/user v0.3.0/go.mod h1:bG+tYYYJgaMtRKgEmuueC0hJEAZWwtIbZTB+85uoHjs=
 github.com/moby/sys/userns v0.1.0 h1:tVLXkFOxVu9A64/yh59slHVv9ahO9UIev4JZusOLG/g=
 github.com/moby/sys/userns v0.1.0/go.mod h1:IHUYgu/kao6N8YZlp9Cf444ySSvCmDlmzUcYfDHOl28=
-github.com/moby/term v0.5.0 h1:xt8Q1nalod/v7BqbG21f8mQPqH+xAaC9C3N3wfWbVP0=
+github.com/moby/term v0.5.2 h1:6qk3FJAFDs6i/q3W/pQ97SX192qKfZgGjCQqfCJkgzQ=
-github.com/moby/term v0.5.0/go.mod h1:8FzsFHVUBGZdbDsJw/ot+X+d5HLUbvklYLJ9uGfcI3Y=
+github.com/moby/term v0.5.2/go.mod h1:d3djjFCrjnB+fl8NJux+EJzu0msscUP+f8it8hPkFLc=
 github.com/modern-go/concurrent v0.0.0-20180228061459-e0a39a4cb421/go.mod h1:6dJC0mAP4ikYIbvyc7fijjWJddQyLn8Ig3JB5CqoB9Q=
 github.com/modern-go/reflect2 v0.0.0-20180701023420-4b7aa43c6742/go.mod h1:bx2lNnkwVCuqBIxFjflWJWanXIb3RllmbCylyMrvgv0=
 github.com/morikuni/aec v1.0.0 h1:nP9CBfwrvYnBRgY6qfDQkygYDmYwOilePFkwzv4dU8A=
@@ -510,8 +518,8 @@ github.com/netbirdio/service v0.0.0-20240911161631-f62744f42502 h1:3tHlFmhTdX9ax
 github.com/netbirdio/service v0.0.0-20240911161631-f62744f42502/go.mod h1:CIMRFEJVL+0DS1a3Nx06NaMn4Dz63Ng6O7dl0qH0zVM=
 github.com/netbirdio/signal-dispatcher/dispatcher v0.0.0-20250805121659-6b4ac470ca45 h1:ujgviVYmx243Ksy7NdSwrdGPSRNE3pb8kEDSpH0QuAQ=
 github.com/netbirdio/signal-dispatcher/dispatcher v0.0.0-20250805121659-6b4ac470ca45/go.mod h1:5/sjFmLb8O96B5737VCqhHyGRzNFIaN/Bu7ZodXc3qQ=
-github.com/netbirdio/wireguard-go v0.0.0-20260523085312-4b4a4e36017f h1:ff2D57RBjWtyQ2wVwJOxOgXAXOe/J2lJWtSX0Bz/BRk=
+github.com/netbirdio/wireguard-go v0.0.0-20260628102922-2834bebf6c1a h1:3CWK+yTvRKOcC0Q8VCTGy4l60TEb27CQVS7LkMxwjmw=
-github.com/netbirdio/wireguard-go v0.0.0-20260523085312-4b4a4e36017f/go.mod h1:rpwXGsirqLqN2L0JDJQlwOboGHmptD5ZD6T2VmcqhTw=
+github.com/netbirdio/wireguard-go v0.0.0-20260628102922-2834bebf6c1a/go.mod h1:rpwXGsirqLqN2L0JDJQlwOboGHmptD5ZD6T2VmcqhTw=
 github.com/nfnt/resize v0.0.0-20180221191011-83c6a9932646 h1:zYyBkD/k9seD2A7fsi6Oo2LfFZAehjjQMERAvZLEDnQ=
 github.com/nfnt/resize v0.0.0-20180221191011-83c6a9932646/go.mod h1:jpp1/29i3P1S/RLdc7JQKbRpFeM1dOBd8T9ki5s+AY8=
 github.com/nicksnyder/go-i18n/v2 v2.5.1 h1:IxtPxYsR9Gp60cGXjfuR/llTqV8aYMsC472zD0D1vHk=
@@ -542,6 +550,8 @@ github.com/opencontainers/go-digest v1.0.0 h1:apOUWs51W5PlhuyGyz9FCeeBIOUDA/6nW8
 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/ory/dockertest/v4 v4.0.0 h1:i19aFsO/VXE0VrMk4ifnKW4G/KIJ93PCjLOslxXoPME=
+github.com/ory/dockertest/v4 v4.0.0/go.mod h1:b5Ofu8VIxWNhXFvQcLu17pRNQdoUBKtXBW74G4Ygzx8=
 github.com/oschwald/maxminddb-golang v1.12.0 h1:9FnTOD0YOhP7DGxGsq4glzpGy5+w7pq50AS6wALUMYs=
 github.com/oschwald/maxminddb-golang v1.12.0/go.mod h1:q0Nob5lTCqyQ8WT6FYgS1L7PXKVVbgiymefNwIjPzgY=
 github.com/patrickmn/go-cache v2.1.0+incompatible h1:HRMgzkcYKYpi3C8ajMPV8OFXaaRUnok+kx1WdO15EQc=
@@ -973,11 +983,13 @@ gorm.io/driver/sqlite v1.5.7/go.mod h1:U+J8craQU6Fzkcvu8oLeAQmi50TkwPEhHDEjQZXDa
 gorm.io/gorm v1.25.7/go.mod h1:hbnx/Oo0ChWMn1BIhpy1oYozzpM15i4YPuHDmfYtwg8=
 gorm.io/gorm v1.25.12 h1:I0u8i2hWQItBq1WfE0o2+WuL9+8L21K9e2HHSTE/0f8=
 gorm.io/gorm v1.25.12/go.mod h1:xh7N7RHfYlNc5EmcI/El95gXusucDrQnHXe0+CgWcLQ=
-gotest.tools/v3 v3.5.1 h1:EENdUnS3pdur5nybKYIh2Vfgc8IUNBjxDPSjtiJcOzU=
+gotest.tools/v3 v3.5.2 h1:7koQfIKdy+I8UTetycgUqXWSDwpgv193Ka+qRsmBY8Q=
-gotest.tools/v3 v3.5.1/go.mod h1:isy3WKz7GK6uNw/sbHzfKBLvlvXwUyV06n6brMxxopU=
+gotest.tools/v3 v3.5.2/go.mod h1:LtdLGcnqToBH83WByAAi/wiwSFCArdFIUV/xxN4pcjA=
 gvisor.dev/gvisor v0.0.0-20260219192049-0f2374377e89 h1:mGJaeA61P8dEHTqdvAgc70ZIV3QoUoJcXCRyyjO26OA=
 gvisor.dev/gvisor v0.0.0-20260219192049-0f2374377e89/go.mod h1:QkHjoMIBaYtpVufgwv3keYAbln78mBoCuShZrPrer1Q=
 howett.net/plist v1.0.1 h1:37GdZ8tP09Q35o9ych3ehygcsL+HqKSwzctveSlarvM=
 howett.net/plist v1.0.1/go.mod h1:lqaXoTrLY4hg8tnEzNru53gicrbv7rrk+2xJA/7hw9g=
+pgregory.net/rapid v1.2.0 h1:keKAYRcjm+e1F0oAuU5F5+YPAWcyxNNRK2wud503Gnk=
+pgregory.net/rapid v1.2.0/go.mod h1:PY5XlDGj0+V1FCq0o192FdRhpKHGTRIWBgqjDBTrq04=
 rsc.io/qr v0.2.0 h1:6vBLea5/NRMVTz8V66gipeLycZMl/+UlFmk8DvqQ6WY=
 rsc.io/qr v0.2.0/go.mod h1:IF+uZjkb9fqyeF/4tlBoynqmQxUoPfWEKh921coOuXs=

infrastructure_files/getting-started.sh

@@ -351,6 +351,11 @@ initialize_default_values() {
   NETBIRD_STUN_PORT=3478
 
   # Docker images
+  # Record whether the operator explicitly pinned the server/proxy images via
+  # env vars, so the agent-network preset can pick its own defaults without
+  # clobbering an explicit override.
+  NETBIRD_SERVER_IMAGE_EXPLICIT=${NETBIRD_SERVER_IMAGE:+true}
+  NETBIRD_PROXY_IMAGE_EXPLICIT=${NETBIRD_PROXY_IMAGE:+true}
   DASHBOARD_IMAGE=${DASHBOARD_IMAGE:-"netbirdio/dashboard:latest"}
   # Combined server replaces separate signal, relay, and management containers
   NETBIRD_SERVER_IMAGE=${NETBIRD_SERVER_IMAGE:-"netbirdio/netbird-server:latest"}
@@ -398,7 +403,53 @@ configure_domain() {
   return 0
 }
 
+apply_agent_network_preset() {
+  # Agent-network turnkey install: built-in Traefik + NetBird Proxy with
+  # NB_PROXY_PRIVATE=true, dashboard locked to agent-network-only mode.
+  # Bypasses every reverse-proxy / proxy / CrowdSec prompt. The only
+  # inputs we still need from the operator are the domain (handled by
+  # configure_domain via NETBIRD_DOMAIN env var or interactive prompt)
+  # and the ACME email — both honor env vars first and fall back to a
+  # prompt only when unset. CrowdSec is intentionally off.
+  REVERSE_PROXY_TYPE="0"
+  ENABLE_PROXY="true"
+  ENABLE_CROWDSEC="false"
+
+  # Agent-network ships dedicated server/proxy images. Honor an explicit
+  # env override; otherwise pin the agent-network builds.
+  if [[ "${NETBIRD_SERVER_IMAGE_EXPLICIT}" != "true" ]]; then
+    NETBIRD_SERVER_IMAGE="netbirdio/netbird-server:0.74.0-rc.2"
+  fi
+  if [[ "${NETBIRD_PROXY_IMAGE_EXPLICIT}" != "true" ]]; then
+    NETBIRD_PROXY_IMAGE="netbirdio/reverse-proxy:0.74.0-rc.2"
+  fi
+
+  if [[ -n "${NETBIRD_LETSENCRYPT_EMAIL}" ]]; then
+    TRAEFIK_ACME_EMAIL="${NETBIRD_LETSENCRYPT_EMAIL}"
+  else
+    TRAEFIK_ACME_EMAIL=$(read_traefik_acme_email)
+  fi
+
+  echo "" > /dev/stderr
+  echo "Agent-network preset enabled (NETBIRD_AGENT_NETWORK=true):" > /dev/stderr
+  echo "  - reverse proxy: built-in Traefik" > /dev/stderr
+  echo "  - NetBird Proxy: enabled with NB_PROXY_PRIVATE=true" > /dev/stderr
+  echo "  - server image: ${NETBIRD_SERVER_IMAGE}" > /dev/stderr
+  echo "  - proxy image: ${NETBIRD_PROXY_IMAGE}" > /dev/stderr
+  echo "  - dashboard: NETBIRD_AGENT_NETWORK_ONLY=true" > /dev/stderr
+  echo "  - CrowdSec: disabled" > /dev/stderr
+  echo "  - Let's Encrypt email: ${TRAEFIK_ACME_EMAIL}" > /dev/stderr
+  echo "" > /dev/stderr
+}
+
 configure_reverse_proxy() {
+  # Short-circuit: agent-network preset locks every reverse-proxy /
+  # proxy / CrowdSec choice and bypasses the interactive prompts.
+  if [[ "${NETBIRD_AGENT_NETWORK}" == "true" ]]; then
+    apply_agent_network_preset
+    return 0
+  fi
+
   # Prompt for reverse proxy type
   REVERSE_PROXY_TYPE=$(read_reverse_proxy_type)
 
@@ -910,6 +961,15 @@ NGINX_SSL_PORT=443
 # Letsencrypt
 LETSENCRYPT_DOMAIN=none
 EOF
+
+  if [[ "${NETBIRD_AGENT_NETWORK}" == "true" ]]; then
+    cat <<EOF
+# Agent-network preset: dashboard hides the standard NetBird surfaces
+# and exposes only the AI Observability + agent-network configuration
+# pages. Paired with NB_PROXY_PRIVATE=true on the proxy side.
+NETBIRD_AGENT_NETWORK_ONLY=true
+EOF
+  fi
   return 0
 }
 
@@ -946,6 +1006,17 @@ NB_PROXY_PROXY_PROTOCOL=true
 NB_PROXY_TRUSTED_PROXIES=$TRAEFIK_IP
 EOF
 
+  if [[ "${NETBIRD_AGENT_NETWORK}" == "true" ]]; then
+    cat <<EOF
+# Agent-network preset: turn the proxy into the private reverse-proxy
+# ingress for agent-network synth services. Disables the public-facing
+# surface so the proxy serves only synth-generated routes (the
+# llm_router-driven LLM endpoints) and the per-account inbound
+# listeners on the embedded netstack.
+NB_PROXY_PRIVATE=true
+EOF
+  fi
+
   if [[ "$ENABLE_CROWDSEC" == "true" && -n "$CROWDSEC_BOUNCER_KEY" ]]; then
     cat <<EOF
 NB_PROXY_CROWDSEC_API_URL=http://crowdsec:8080
@@ -1326,12 +1397,20 @@ print_builtin_traefik_instructions() {
     echo "  - 51820/udp (WIREGUARD - (optional) for P2P proxy connections)"
   fi
   echo ""
-  echo "This setup is ideal for homelabs and smaller organization deployments."
+  if [[ "${NETBIRD_AGENT_NETWORK}" == "true" ]]; then
-  echo "For enterprise environments requiring high availability and advanced integrations,"
+    echo "For enterprise environments requiring high availability and advanced integrations,"
-  echo "consider a commercial on-prem license or scaling your open source deployment:"
+    echo "consider a commercial on-prem license:"
-  echo ""
+    echo ""
-  echo "  Commercial license: https://netbird.io/pricing#on-prem"
+    echo "  Commercial license: https://netbird.ai/pricing"
-  echo "  Scaling guide:      https://docs.netbird.io/scaling-your-self-hosted-deployment"
+    echo "  Documentation: https://docs.netbird.io/agent-network"
+  else
+    echo "This setup is ideal for homelabs and smaller organization deployments."
+    echo "For enterprise environments requiring high availability and advanced integrations,"
+    echo "consider a commercial on-prem license or scaling your open source deployment:"
+    echo ""
+    echo "  Commercial license: https://netbird.io/pricing#on-prem"
+    echo "  Scaling guide:      https://docs.netbird.io/scaling-your-self-hosted-deployment"
+  fi
   echo ""
   if [[ "$ENABLE_PROXY" == "true" ]]; then
     echo "NetBird Proxy:"
@@ -1354,6 +1433,11 @@ print_builtin_traefik_instructions() {
       echo ""
     fi
   fi
+  if [[ "${NETBIRD_AGENT_NETWORK}" == "true" ]]; then
+    echo "Note: The public domain is only for setting up secure connections."
+    echo "Your APIs and agent services remain private and are never exposed publicly."
+    echo ""
+  fi
   return 0
 }
 

management/internals/controllers/network_map/controller/controller.go

@@ -497,7 +497,7 @@ func (c *Controller) BufferUpdateAffectedPeers(ctx context.Context, accountID st
 		c.accountManagerMetrics.CountUpdateAccountPeersTriggered(string(reason.Resource), string(reason.Operation))
 	}
 
-	log.WithContext(ctx).Tracef("buffer updating %d affected peers for account %s from %s", len(peerIDs), accountID, util.GetCallerName())
+	log.WithContext(ctx).Tracef("buffer updating %d affected peers for account %s from %s with reason %s/%s", len(peerIDs), accountID, util.GetCallerName(), reason.Operation, reason.Resource)
 
 	bufUpd, _ := c.affectedPeerUpdateLocks.LoadOrStore(accountID, &bufferAffectedUpdate{
 		peerIDs: make(map[string]struct{}),
@@ -610,12 +610,10 @@ func (c *Controller) GetValidatedPeerWithMap(ctx context.Context, isRequiresAppr
 		return nil, nil, 0, err
 	}
 
-	startPosture := time.Now()
 	postureChecks, err := c.getPeerPostureChecks(account, peerID)
 	if err != nil {
 		return nil, nil, 0, err
 	}
-	log.WithContext(ctx).Debugf("getPeerPostureChecks took %s", time.Since(startPosture))
 
 	accountZones, err := c.repo.GetAccountZones(ctx, account.Id)
 	if err != nil {

management/internals/shared/grpc/loginfilter.go

@@ -13,7 +13,7 @@ const (
 	reconnThreshold   = 5 * time.Minute
 	baseBlockDuration = 10 * time.Minute // Duration for which a peer is banned after exceeding the reconnection limit
 	reconnLimitForBan = 30               // Number of reconnections within the reconnTreshold that triggers a ban
-	metaChangeLimit   = 3                // Number of reconnections with different metadata that triggers a ban of one peer
+	metaChangeLimit   = 5                // Number of reconnections with different metadata that triggers a ban of one peer
 )
 
 type lfConfig struct {
@@ -139,7 +139,7 @@ func (l *loginFilter) addLogin(wgPubKey string, metaHash uint64) {
 	state.lastSeen = now
 }
 
-func metaHash(meta nbpeer.PeerSystemMeta, pubip string) uint64 {
+func metaHash(meta nbpeer.PeerSystemMeta) uint64 {
 	h := fnv.New64a()
 
 	h.Write([]byte(meta.WtVersion))
@@ -147,14 +147,6 @@ func metaHash(meta nbpeer.PeerSystemMeta, pubip string) uint64 {
 	h.Write([]byte(meta.KernelVersion))
 	h.Write([]byte(meta.Hostname))
 	h.Write([]byte(meta.SystemSerialNumber))
-	h.Write([]byte(pubip))
 
-	macs := uint64(0)
+	return h.Sum64()
-	for _, na := range meta.NetworkAddresses {
-		for _, r := range na.Mac {
-			macs += uint64(r)
-		}
-	}
-
-	return h.Sum64() + macs
 }

management/internals/shared/grpc/loginfilter_test.go

@@ -164,9 +164,7 @@ func BenchmarkHashingMethods(b *testing.B) {
 		KernelVersion:      "5.15.0-76-generic",
 		Hostname:           "prod-server-database-01",
 		SystemSerialNumber: "PC-1234567890",
-		NetworkAddresses:   []nbpeer.NetworkAddress{{Mac: "00:1B:44:11:3A:B7"}, {Mac: "00:1B:44:11:3A:B8"}},
 	}
-	pubip := "8.8.8.8"
 
 	var resultString string
 	var resultUint uint64
@@ -175,7 +173,7 @@ func BenchmarkHashingMethods(b *testing.B) {
 		b.ReportAllocs()
 		b.ResetTimer()
 		for i := 0; i < b.N; i++ {
-			resultString = builderString(meta, pubip)
+			resultString = builderString(meta)
 		}
 	})
 
@@ -183,7 +181,7 @@ func BenchmarkHashingMethods(b *testing.B) {
 		b.ReportAllocs()
 		b.ResetTimer()
 		for i := 0; i < b.N; i++ {
-			resultString = fnvHashToString(meta, pubip)
+			resultString = fnvHashToString(meta)
 		}
 	})
 
@@ -191,7 +189,7 @@ func BenchmarkHashingMethods(b *testing.B) {
 		b.ReportAllocs()
 		b.ResetTimer()
 		for i := 0; i < b.N; i++ {
-			resultUint = metaHash(meta, pubip)
+			resultUint = metaHash(meta)
 		}
 	})
 
@@ -199,29 +197,20 @@ func BenchmarkHashingMethods(b *testing.B) {
 	_ = resultUint
 }
 
-func fnvHashToString(meta nbpeer.PeerSystemMeta, pubip string) string {
+func fnvHashToString(meta nbpeer.PeerSystemMeta) string {
 	h := fnv.New64a()
 
-	if len(meta.NetworkAddresses) != 0 {
-		for _, na := range meta.NetworkAddresses {
-			h.Write([]byte(na.Mac))
-		}
-	}
-
 	h.Write([]byte(meta.WtVersion))
 	h.Write([]byte(meta.OSVersion))
 	h.Write([]byte(meta.KernelVersion))
 	h.Write([]byte(meta.Hostname))
 	h.Write([]byte(meta.SystemSerialNumber))
-	h.Write([]byte(pubip))
 
 	return strconv.FormatUint(h.Sum64(), 16)
 }
 
-func builderString(meta nbpeer.PeerSystemMeta, pubip string) string {
+func builderString(meta nbpeer.PeerSystemMeta) string {
-	mac := getMacAddress(meta.NetworkAddresses)
+	estimatedSize := len(meta.WtVersion) + len(meta.OSVersion) + len(meta.KernelVersion) + len(meta.Hostname) + len(meta.SystemSerialNumber) + 4
-	estimatedSize := len(meta.WtVersion) + len(meta.OSVersion) + len(meta.KernelVersion) + len(meta.Hostname) + len(meta.SystemSerialNumber) +
-		len(pubip) + len(mac) + 6
 
 	var b strings.Builder
 	b.Grow(estimatedSize)
@@ -235,23 +224,10 @@ func builderString(meta nbpeer.PeerSystemMeta, pubip string) string {
 	b.WriteString(meta.Hostname)
 	b.WriteByte('|')
 	b.WriteString(meta.SystemSerialNumber)
-	b.WriteByte('|')
-	b.WriteString(pubip)
 
 	return b.String()
 }
 
-func getMacAddress(nas []nbpeer.NetworkAddress) string {
-	if len(nas) == 0 {
-		return ""
-	}
-	macs := make([]string, 0, len(nas))
-	for _, na := range nas {
-		macs = append(macs, na.Mac)
-	}
-	return strings.Join(macs, "/")
-}
-
 func BenchmarkLoginFilter_ParallelLoad(b *testing.B) {
 	filter := newLoginFilterWithCfg(testAdvancedCfg())
 	numKeys := 100000

management/internals/shared/grpc/server.go

@@ -254,7 +254,7 @@ func (s *Server) Sync(req *proto.EncryptedMessage, srv proto.ManagementService_S
 		return mapError(ctx, err)
 	}
 
-	metahashed := metaHash(peerMeta, sRealIP)
+	metahashed := metaHash(peerMeta)
 	if userID == "" && !s.loginFilter.allowLogin(peerKey.String(), metahashed) {
 		if s.appMetrics != nil {
 			s.appMetrics.GRPCMetrics().CountSyncRequestBlocked()
@@ -306,7 +306,7 @@ func (s *Server) Sync(req *proto.EncryptedMessage, srv proto.ManagementService_S
 		log.WithContext(ctx).Tracef("peer system meta has to be provided on sync. Peer %s, remote addr %s", peerKey.String(), realIP)
 	}
 
-	metahash := metaHash(peerMeta, realIP.String())
+	metahash := metaHash(peerMeta)
 	s.loginFilter.addLogin(peerKey.String(), metahash)
 
 	peer, netMap, postureChecks, dnsFwdPort, err := s.accountManager.SyncAndMarkPeer(ctx, accountID, peerKey.String(), peerMeta, realIP, syncStart)
@@ -732,7 +732,7 @@ func (s *Server) Login(ctx context.Context, req *proto.EncryptedMessage) (*proto
 	}
 
 	peerMeta := extractPeerMeta(ctx, loginReq.GetMeta())
-	metahashed := metaHash(peerMeta, sRealIP)
+	metahashed := metaHash(peerMeta)
 	if !s.loginFilter.allowLogin(peerKey.String(), metahashed) {
 		if s.logBlockedPeers {
 			log.WithContext(ctx).Tracef("peer %s with meta hash %d is blocked from login", peerKey.String(), metahashed)
@@ -788,7 +788,11 @@ func (s *Server) Login(ctx context.Context, req *proto.EncryptedMessage) (*proto
 		ExtraDNSLabels:  loginReq.GetDnsLabels(),
 	})
 	if err != nil {
-		log.WithContext(ctx).Warnf("failed logging in peer %s: %s", peerKey, err)
+		if errors.Is(err, internalStatus.ErrNoAuthMethodProvided) {
+			log.WithContext(ctx).Tracef("failed logging in peer %s: %s", peerKey, err)
+		} else {
+			log.WithContext(ctx).Warnf("failed logging in peer %s: %s", peerKey, err)
+		}
 		return nil, mapError(ctx, err)
 	}
 

management/server/affected_peers_coverage_test.go

@@ -41,7 +41,7 @@ func TestAffectedPeers_DependencyCoverageMatrix(t *testing.T) {
 				_, err := s.manager.SavePolicy(ctx, s.accountID, userID, peerToResourcePolicyByGroup(s.sourceGroupID, s.resourceGroupID), true)
 				require.NoError(t, err)
 				return affectedpeers.Change{ChangedPeerIDs: []string{s.routerPeerID}},
-					[]string{s.sourcePeerID}, []string{s.unrelatedPeerID}
+					[]string{s.sourcePeerID, s.routerPeerID}, []string{s.unrelatedPeerID}
 			},
 		},
 		{
@@ -106,11 +106,9 @@ func TestAffectedPeers_DependencyCoverageMatrix(t *testing.T) {
 			change, mustContain, mustExclude := r.build(t, s, ctx)
 			affected := resolveAffected(t, s.manager.Store, s.accountID, change)
 
-			for _, id := range mustContain {
+			assert.ElementsMatch(t, affected, mustContain, "expected peer to be affected")
-				assert.Contains(t, affected, id, "expected peer to be affected")
+			for _, peerID := range mustExclude {
-			}
+				assert.NotContains(t, affected, peerID, "peer must not be affected")
-			for _, id := range mustExclude {
-				assert.NotContains(t, affected, id, "peer must not be affected")
 			}
 		})
 	}

management/server/affected_peers_router_paths_test.go

@@ -251,7 +251,9 @@ func TestAffectedPeers_E2E_UpdateResource_DestinationResourcePolicy_RefreshesSou
 	}
 }
 
-func TestAffectedPeers_E2E_UpdateResource_DisabledSiblingRouter_StillBridged(t *testing.T) {
+// A disabled sibling router routes to nobody, so updating a resource on its network
+// must NOT refresh its peer (the enabled router carries the bridge instead).
+func TestAffectedPeers_E2E_UpdateResource_DisabledSiblingRouterNotBridged(t *testing.T) {
 	s := setupRouterScenario(t, true)
 	ctx := context.Background()
 
@@ -274,13 +276,18 @@ func TestAffectedPeers_E2E_UpdateResource_DisabledSiblingRouter_StillBridged(t *
 	require.NoError(t, err)
 
 	disabledCh := s.updateManager.CreateChannel(ctx, disabledRouterPeer.ID)
-	t.Cleanup(func() { s.updateManager.CloseChannel(ctx, disabledRouterPeer.ID) })
+	enabledCh := s.updateManager.CreateChannel(ctx, s.routerPeerID)
+	t.Cleanup(func() {
+		s.updateManager.CloseChannel(ctx, disabledRouterPeer.ID)
+		s.updateManager.CloseChannel(ctx, s.routerPeerID)
+	})
 
-	settleAffectedUpdates(disabledCh)
+	settleAffectedUpdates(disabledCh, enabledCh)
 
 	done := make(chan struct{})
 	go func() {
-		peerShouldReceiveUpdate(t, disabledCh)
+		peerShouldReceiveUpdate(t, enabledCh)
+		peerShouldNotReceiveUpdate(t, disabledCh)
 		close(done)
 	}()
 
@@ -298,7 +305,7 @@ func TestAffectedPeers_E2E_UpdateResource_DisabledSiblingRouter_StillBridged(t *
 	select {
 	case <-done:
 	case <-time.After(peerUpdateTimeout):
-		t.Error("timeout: resource update did not refresh the disabled sibling router's peer")
+		t.Error("timeout")
 	}
 }
 

management/server/affected_peers_router_test.go

@@ -682,6 +682,9 @@ func TestAffectedPeers_AllRoutingPeers_Network(t *testing.T) {
 	assert.Contains(t, affected, secondRouterPeer.ID, "second routing peer on the same network must also be affected")
 }
 
+// A disabled router in the snapshot routes to nobody, so it is skipped when the
+// walk scans existing account data: a policy edit still folds the literal source
+// group, but not the disabled router's peer.
 func TestAffectedPeers_DisabledRouter(t *testing.T) {
 	s := setupRouterScenario(t, true)
 	ctx := context.Background()
@@ -694,11 +697,13 @@ func TestAffectedPeers_DisabledRouter(t *testing.T) {
 
 	affected := s.resolvePolicyAffected(ctx, peerToResourcePolicyByGroup(s.sourceGroupID, s.resourceGroupID))
 
-	assert.Contains(t, affected, s.sourcePeerID, "source peer must be affected")
+	assert.Contains(t, affected, s.sourcePeerID, "source peer (literal policy source group) must be affected")
-	assert.Contains(t, affected, s.routerPeerID,
+	assert.NotContains(t, affected, s.routerPeerID,
-		"disabled router's peer must still be affected: Enabled must not gate affected-peers")
+		"a disabled router routes to nobody, so its peer must not be folded from snapshot data")
 }
 
+// A disabled resource in the snapshot is skipped: the policy edit still folds the
+// literal source group, but the resource no longer bridges to its network's router.
 func TestAffectedPeers_DisabledResource(t *testing.T) {
 	s := setupRouterScenario(t, true)
 	ctx := context.Background()
@@ -710,9 +715,9 @@ func TestAffectedPeers_DisabledResource(t *testing.T) {
 
 	affected := s.resolvePolicyAffected(ctx, peerToResourcePolicyByGroup(s.sourceGroupID, s.resourceGroupID))
 
-	assert.Contains(t, affected, s.sourcePeerID, "source peer must be affected")
+	assert.Contains(t, affected, s.sourcePeerID, "source peer (literal policy source group) must be affected")
-	assert.Contains(t, affected, s.routerPeerID,
+	assert.NotContains(t, affected, s.routerPeerID,
-		"disabled resource must still resolve the routing peer: Enabled must not gate affected-peers")
+		"a disabled resource routes to nobody, so its network's router must not be folded from snapshot data")
 }
 
 func TestAffectedPeers_DisabledRule(t *testing.T) {

management/server/affected_peers_test.go

@@ -96,33 +96,54 @@ func affectedGroupID(i int) string   { return fmt.Sprintf("affected-grp-%d", i)
 func affectedGroupName(i int) string { return fmt.Sprintf("AffectedGroup%d", i) }
 
 func TestCollectGroupChange_PolicyLinked(t *testing.T) {
-	manager, s, accountID, _, groupIDs := setupAffectedPeersTest(t)
+	manager, s, accountID, peerIDs, groupIDs := setupAffectedPeersTest(t)
 	ctx := context.Background()
 
 	_, err := manager.SavePolicy(ctx, accountID, userID, &types.Policy{
 		Enabled: true,
 		Rules: []*types.PolicyRule{
 			{
-				Enabled:       true,
+				Enabled:             true,
-				Sources:       []string{groupIDs[0]},
+				Sources:             []string{groupIDs[0]},
-				Destinations:  []string{groupIDs[1]},
+				Destinations:        []string{groupIDs[1]},
-				Bidirectional: true,
+				SourceResource:      types.Resource{ID: peerIDs[0], Type: types.ResourceTypePeer},
-				Action:        types.PolicyTrafficActionAccept,
+				DestinationResource: types.Resource{ID: peerIDs[1], Type: types.ResourceTypePeer},
+				Bidirectional:       true,
+				Action:              types.PolicyTrafficActionAccept,
+			},
+			{
+				Enabled:             true,
+				Sources:             []string{groupIDs[0]},
+				Destinations:        []string{groupIDs[1]},
+				SourceResource:      types.Resource{ID: peerIDs[2], Type: types.ResourceTypeHost},
+				DestinationResource: types.Resource{ID: peerIDs[3], Type: types.ResourceTypeHost},
+				Bidirectional:       true,
+				Action:              types.PolicyTrafficActionAccept,
+			},
+			{
+				Enabled:             true,
+				Sources:             []string{groupIDs[0]},
+				Destinations:        []string{groupIDs[1]},
+				SourceResource:      types.Resource{ID: "", Type: types.ResourceTypePeer},
+				DestinationResource: types.Resource{ID: "", Type: types.ResourceTypePeer},
+				Bidirectional:       true,
+				Action:              types.PolicyTrafficActionAccept,
 			},
 		},
 	}, true)
 	require.NoError(t, err)
 
-	groups, _ := collectGroupChangeAffectedGroups(ctx, s, accountID, []string{groupIDs[0]})
+	groups, directPeers := collectGroupChangeAffectedGroups(ctx, s, accountID, []string{groupIDs[0]})
-	assert.Contains(t, groups, groupIDs[0])
+	assert.ElementsMatch(t, groups, []string{groupIDs[0], groupIDs[1]})
-	assert.Contains(t, groups, groupIDs[1])
+	assert.ElementsMatch(t, directPeers, []string{peerIDs[1]})
 
-	groups, _ = collectGroupChangeAffectedGroups(ctx, s, accountID, []string{groupIDs[1]})
+	groups, directPeers = collectGroupChangeAffectedGroups(ctx, s, accountID, []string{groupIDs[1]})
-	assert.Contains(t, groups, groupIDs[0])
+	assert.ElementsMatch(t, groups, []string{groupIDs[0], groupIDs[1]})
-	assert.Contains(t, groups, groupIDs[1])
+	assert.ElementsMatch(t, directPeers, []string{peerIDs[0]})
 
-	groups, _ = collectGroupChangeAffectedGroups(ctx, s, accountID, []string{groupIDs[2]})
+	groups, directPeers = collectGroupChangeAffectedGroups(ctx, s, accountID, []string{groupIDs[2]})
 	assert.Empty(t, groups)
+	assert.Empty(t, directPeers)
 }
 
 func TestCollectGroupChange_PolicyWithDirectPeerResource(t *testing.T) {
@@ -133,20 +154,44 @@ func TestCollectGroupChange_PolicyWithDirectPeerResource(t *testing.T) {
 		Enabled: true,
 		Rules: []*types.PolicyRule{
 			{
-				Enabled:        true,
+				Enabled:             true,
-				Sources:        []string{groupIDs[0]},
+				Sources:             []string{groupIDs[0]},
-				SourceResource: types.Resource{ID: peerIDs[3], Type: types.ResourceTypePeer},
+				SourceResource:      types.Resource{ID: peerIDs[3], Type: types.ResourceTypePeer},
-				Destinations:   []string{groupIDs[1]},
+				DestinationResource: types.Resource{ID: peerIDs[4], Type: types.ResourceTypePeer},
-				Action:         types.PolicyTrafficActionAccept,
+				Destinations:        []string{groupIDs[1]},
+				Action:              types.PolicyTrafficActionAccept,
+			},
+			{
+				Enabled:             true,
+				Sources:             []string{groupIDs[0]},
+				SourceResource:      types.Resource{ID: peerIDs[1], Type: types.ResourceTypeHost},
+				DestinationResource: types.Resource{ID: peerIDs[2], Type: types.ResourceTypeHost},
+				Destinations:        []string{groupIDs[1]},
+				Action:              types.PolicyTrafficActionAccept,
+			},
+			{
+				Enabled:             true,
+				Sources:             []string{groupIDs[0]},
+				SourceResource:      types.Resource{ID: "", Type: types.ResourceTypePeer},
+				DestinationResource: types.Resource{ID: "", Type: types.ResourceTypePeer},
+				Destinations:        []string{groupIDs[1]},
+				Action:              types.PolicyTrafficActionAccept,
 			},
 		},
 	}, true)
 	require.NoError(t, err)
 
 	groups, directPeers := collectGroupChangeAffectedGroups(ctx, s, accountID, []string{groupIDs[0]})
-	assert.Contains(t, groups, groupIDs[0])
+	assert.ElementsMatch(t, groups, []string{groupIDs[0], groupIDs[1]})
-	assert.Contains(t, groups, groupIDs[1])
+	assert.ElementsMatch(t, directPeers, []string{peerIDs[4]})
-	assert.Contains(t, directPeers, peerIDs[3])
+
+	groups, directPeers = collectGroupChangeAffectedGroups(ctx, s, accountID, []string{groupIDs[1]})
+	assert.ElementsMatch(t, groups, []string{groupIDs[0], groupIDs[1]})
+	assert.ElementsMatch(t, directPeers, []string{peerIDs[3]})
+
+	groups, directPeers = collectGroupChangeAffectedGroups(ctx, s, accountID, []string{groupIDs[2]})
+	assert.Empty(t, groups)
+	assert.Empty(t, directPeers)
 }
 
 func TestCollectGroupChange_PolicyWithNonPeerResource_NoDirectPeers(t *testing.T) {
@@ -168,8 +213,7 @@ func TestCollectGroupChange_PolicyWithNonPeerResource_NoDirectPeers(t *testing.T
 	require.NoError(t, err)
 
 	groups, directPeers := collectGroupChangeAffectedGroups(ctx, s, accountID, []string{groupIDs[0]})
-	assert.Contains(t, groups, groupIDs[0])
+	assert.ElementsMatch(t, groups, []string{groupIDs[0], groupIDs[1]})
-	assert.Contains(t, groups, groupIDs[1])
 	assert.Empty(t, directPeers, "non-peer resources should not produce direct peer IDs")
 }
 
@@ -294,6 +338,7 @@ func TestCollectGroupChange_NetworkRouterLinked(t *testing.T) {
 		AccountID:  accountID,
 		PeerGroups: []string{groupIDs[0]},
 		Peer:       peerIDs[3],
+		Enabled:    true,
 	})
 	require.NoError(t, err)
 
@@ -324,6 +369,7 @@ func TestCollectGroupChange_NetworkRouterPeerOnlyNoGroups(t *testing.T) {
 		NetworkID: net1.ID,
 		AccountID: accountID,
 		Peer:      peerIDs[4],
+		Enabled:   true,
 	})
 	require.NoError(t, err)
 
@@ -373,17 +419,11 @@ func TestCollectGroupChange_MultipleEntities(t *testing.T) {
 	require.NoError(t, err)
 
 	groups, directPeers := collectGroupChangeAffectedGroups(ctx, s, accountID, []string{groupIDs[0]})
-	assert.Contains(t, groups, groupIDs[0])
+	assert.ElementsMatch(t, groups, []string{groupIDs[0], groupIDs[1]})
-	assert.Contains(t, groups, groupIDs[1])
-	assert.NotContains(t, groups, groupIDs[2])
-	assert.NotContains(t, groups, groupIDs[3])
 	assert.Empty(t, directPeers)
 
 	groups, directPeers = collectGroupChangeAffectedGroups(ctx, s, accountID, []string{groupIDs[3]})
-	assert.Contains(t, groups, groupIDs[2])
+	assert.ElementsMatch(t, groups, []string{groupIDs[2], groupIDs[3]})
-	assert.Contains(t, groups, groupIDs[3])
-	assert.NotContains(t, groups, groupIDs[0])
-	assert.NotContains(t, groups, groupIDs[1])
 	assert.Empty(t, directPeers)
 }
 
@@ -452,8 +492,9 @@ func TestResolveAffectedPeers_PolicyBetweenTwoGroups(t *testing.T) {
 	result = manager.resolveAffectedPeersForPeerChanges(ctx, s, accountID, []string{peerIDs[1]})
 	assert.ElementsMatch(t, []string{peerIDs[0], peerIDs[1]}, result)
 
+	// peerIDs[2] is unrelated to the route; only its own map can change.
 	result = manager.resolveAffectedPeersForPeerChanges(ctx, s, accountID, []string{peerIDs[2]})
-	assert.Empty(t, result)
+	assert.ElementsMatch(t, []string{peerIDs[2]}, result)
 }
 
 func TestResolveAffectedPeers_PolicyThreeGroups(t *testing.T) {
@@ -474,7 +515,7 @@ func TestResolveAffectedPeers_PolicyThreeGroups(t *testing.T) {
 	require.NoError(t, err)
 
 	result := manager.resolveAffectedPeersForPeerChanges(ctx, s, accountID, []string{peerIDs[0]})
-	assert.ElementsMatch(t, []string{peerIDs[0], peerIDs[1], peerIDs[2]}, result)
+	assert.ElementsMatch(t, []string{peerIDs[0], peerIDs[2]}, result)
 }
 
 func TestResolveAffectedPeers_RoutePeerGroups(t *testing.T) {
@@ -506,8 +547,9 @@ func TestResolveAffectedPeers_RoutePeerGroups(t *testing.T) {
 	result = manager.resolveAffectedPeersForPeerChanges(ctx, s, accountID, []string{peerIDs[1]})
 	assert.ElementsMatch(t, []string{peerIDs[0], peerIDs[1]}, result)
 
+	// peerIDs[2] is in no policy; only its own map can change, so it refreshes itself.
 	result = manager.resolveAffectedPeersForPeerChanges(ctx, s, accountID, []string{peerIDs[2]})
-	assert.Empty(t, result)
+	assert.ElementsMatch(t, []string{peerIDs[2]}, result)
 }
 
 func TestResolveAffectedPeers_RouteWithDirectPeer(t *testing.T) {
@@ -564,9 +606,9 @@ func TestResolveAffectedPeers_RouteWithAccessControlGroups(t *testing.T) {
 	result := manager.resolveAffectedPeersForPeerChanges(ctx, s, accountID, []string{peerIDs[2]})
 	assert.ElementsMatch(t, []string{peerIDs[0], peerIDs[1], peerIDs[2]}, result)
 
-	// peer3 is unrelated
+	// peer3 is unrelated to the route; only its own map can change.
 	result = manager.resolveAffectedPeersForPeerChanges(ctx, s, accountID, []string{peerIDs[3]})
-	assert.Empty(t, result)
+	assert.ElementsMatch(t, []string{peerIDs[3]}, result)
 }
 
 func TestResolveAffectedPeers_NetworkRouter(t *testing.T) {
@@ -587,6 +629,7 @@ func TestResolveAffectedPeers_NetworkRouter(t *testing.T) {
 		AccountID:  accountID,
 		PeerGroups: []string{groupIDs[0]},
 		Peer:       peerIDs[3],
+		Enabled:    true,
 	})
 	require.NoError(t, err)
 
@@ -659,9 +702,13 @@ func TestResolveAffectedPeers_PeerInMultipleGroups(t *testing.T) {
 	}, true)
 	require.NoError(t, err)
 
-	// peer0 is in group0 AND group1, so both policies apply
+	// peer0 is in group0 AND group1, so both policies apply. A peer change folds
+	// only the changed peer plus the opposite side of each rule: group2 (peer2) via
+	// the group0 policy and group3 (peer3) via the group1 policy. peer1, a co-member
+	// of group1, is a sibling of the changed peer and must NOT refresh.
 	result := manager.resolveAffectedPeersForPeerChanges(ctx, s, accountID, []string{peerIDs[0]})
-	assert.ElementsMatch(t, []string{peerIDs[0], peerIDs[1], peerIDs[2], peerIDs[3]}, result)
+	assert.ElementsMatch(t, []string{peerIDs[0], peerIDs[2], peerIDs[3]}, result)
+	assert.NotContains(t, result, peerIDs[1], "co-member of the changed peer's group must not refresh")
 }
 
 func TestResolveAffectedPeers_MultipleChangedPeers(t *testing.T) {
@@ -697,7 +744,7 @@ func TestResolveAffectedPeers_MultipleChangedPeers(t *testing.T) {
 	require.NoError(t, err)
 
 	result := manager.resolveAffectedPeersForPeerChanges(ctx, s, accountID, []string{peerIDs[0], peerIDs[2]})
-	assert.ElementsMatch(t, []string{peerIDs[0], peerIDs[1], peerIDs[2], peerIDs[3]}, result)
+	assert.ElementsMatch(t, []string{peerIDs[0], peerIDs[2], peerIDs[1], peerIDs[3]}, result)
 }
 
 func TestResolveAffectedPeers_SharedGroupAcrossPolicyAndRoute(t *testing.T) {
@@ -854,8 +901,9 @@ func TestAffectedPeers_IsolatedPolicies(t *testing.T) {
 	assert.NotContains(t, result, peerIDs[0])
 	assert.NotContains(t, result, peerIDs[1])
 
+	// peerIDs[4] is in neither isolated policy; only its own map can change.
 	result = manager.resolveAffectedPeersForPeerChanges(ctx, s, accountID, []string{peerIDs[4]})
-	assert.Empty(t, result)
+	assert.ElementsMatch(t, []string{peerIDs[4]}, result)
 }
 
 func TestAffectedPeers_IsolatedRouteAndPolicy(t *testing.T) {
@@ -977,12 +1025,13 @@ func TestAffectedPeers_GroupUpdateOnlyAffectsLinkedPeers(t *testing.T) {
 	})
 }
 
-func TestAffectedPeers_UnlinkedGroupChange_NoUpdates(t *testing.T) {
+// A peer in no policy/route refreshes only itself — no other peer is affected.
+func TestAffectedPeers_UnlinkedPeerChange_RefreshesSelfOnly(t *testing.T) {
 	manager, s, accountID, peerIDs, _ := setupAffectedPeersTest(t)
 	ctx := context.Background()
 
 	result := manager.resolveAffectedPeersForPeerChanges(ctx, s, accountID, []string{peerIDs[0]})
-	assert.Empty(t, result)
+	assert.ElementsMatch(t, []string{peerIDs[0]}, result)
 }
 
 // TestAffectedPeers_PolicyChange_UnrelatedPeerNoUpdate verifies that creating/deleting a
@@ -1332,6 +1381,7 @@ func TestAffectedPeers_NetworkRouterUnlinkedPeerNoUpdate(t *testing.T) {
 		NetworkID:  net1.ID,
 		AccountID:  accountID,
 		PeerGroups: []string{"nr-grpA"},
+		Enabled:    true,
 	})
 	require.NoError(t, err)
 
@@ -1755,7 +1805,9 @@ func TestCollectAffectedFromProxyServices_GroupContainingTargetPeerChanged(t *te
 	assert.Contains(t, directPeers, peerIDs[1], "target peer must be refreshed")
 }
 
-func TestCollectAffectedFromProxyServices_DisabledServiceStillMatches(t *testing.T) {
+// A disabled service in the snapshot proxies nothing, so it is skipped: a changed
+// target peer does not pull in the service's proxy peer.
+func TestCollectAffectedFromProxyServices_DisabledServiceSkipped(t *testing.T) {
 	manager, s, accountID, peerIDs, _ := setupAffectedPeersTest(t)
 	ctx := context.Background()
 
@@ -1781,8 +1833,7 @@ func TestCollectAffectedFromProxyServices_DisabledServiceStillMatches(t *testing
 	require.NoError(t, s.CreateService(ctx, svc))
 
 	_, directPeers := collectPeerChangeAffectedGroups(ctx, manager.Store, accountID, nil, []string{peerIDs[1]})
-	assert.Contains(t, directPeers, peerIDs[0], "disabled service should still trigger a refresh so peers are ready when re-enabled")
+	assert.NotContains(t, directPeers, peerIDs[0], "a disabled service proxies nothing, so its proxy peer must not be folded")
-	assert.Contains(t, directPeers, peerIDs[1], "disabled target should still trigger a refresh")
 }
 
 func TestCollectAffectedFromProxyServices_NonPeerTargetType(t *testing.T) {

management/server/affectedpeers/resolver.go

@@ -6,7 +6,12 @@
 //     and before a delete/removal severs the old state).
 //   - Snapshot.Expand: in-memory walk, no store access. Run AFTER the tx commits.
 //
-// Enabled is never consulted: toggling it is itself an observable change.
+// Enabled handling differs by source. Disabled objects in the SNAPSHOT (existing
+// account policies/resources/routers/routes/proxy services and their rules/targets)
+// route to nobody and are skipped — they cannot affect any peer's map. Objects in
+// the CHANGE itself are processed regardless of Enabled, so disabling one still
+// refreshes the peers that lose access (the toggle is the observable change, and the
+// update carries the old∪new state).
 package affectedpeers
 
 import (
@@ -61,7 +66,8 @@ func Load(ctx context.Context, s store.Store, accountID string, c Change) (*Snap
 // loadCollections reads the policy/route/nameserver/dns/router/resource/proxy
 // collections a Change can touch, gated to what the walk needs.
 func (snap *Snapshot) loadCollections(ctx context.Context, s store.Store, accountID string, c Change) error {
-	hasGroupOrPeerChange := len(c.ChangedGroupIDs) > 0 || len(c.ChangedPeerIDs) > 0 || len(c.Resources) > 0
+	// LinkGroups drive the same policy/route/dns walk as a changed group or peer.
+	hasGroupOrPeerChange := len(c.ChangedGroupIDs) > 0 || len(c.ChangedPeerIDs) > 0 || len(c.LinkGroups) > 0 || len(c.Resources) > 0
 	hasNetworkObject := len(c.Routers) > 0 || len(c.Resources) > 0 || len(c.Networks) > 0
 	// the resource<->router bridge can fire for any of these
 	needsRoutersResources := hasGroupOrPeerChange || len(c.PostureCheckIDs) > 0 || len(c.Policies) > 0 || hasNetworkObject
@@ -76,7 +82,7 @@ func (snap *Snapshot) loadCollections(ctx context.Context, s store.Store, accoun
 			return err
 		}
 	}
-	if len(c.ChangedGroupIDs) > 0 || len(c.ChangedPeerIDs) > 0 {
+	if len(c.ChangedGroupIDs) > 0 || len(c.ChangedPeerIDs) > 0 || len(c.LinkGroups) > 0 {
 		if err := snap.loadDNS(ctx, s, accountID); err != nil {
 			return err
 		}
@@ -174,6 +180,24 @@ type Change struct {
 	// folded in — but only when the group is linked (an unlinked group has no map
 	// impact), matching how current members are handled.
 	RemovedPeersByGroup map[string][]string
+
+	// OutputPeerIDs are peers folded straight into the result without seeding their
+	// group memberships into the walk. Use for the peer whose group membership changed:
+	// the peer itself must refresh, but its OTHER groups did not change, so they must
+	// not be walked. Contrast ChangedPeerIDs, which seeds ALL of the peer's groups
+	// (correct when the peer's own attributes changed, e.g. IP/status).
+	OutputPeerIDs []string
+
+	// LinkGroups are groups used ONLY to match policies/routes/routers and walk to the
+	// OPPOSITE side — they are never expanded to their own members. Use this when a
+	// peer's group membership changed: pass the peer in ChangedPeerIDs and its
+	// group(s) here. The opposite side of the policies the group participates in
+	// refreshes, but the group's other members (siblings) do not — nothing changed for
+	// them. For an intra-group policy (A→A) the opposite side IS the group, so its
+	// members still refresh via the opposite-side fold, exactly when they genuinely
+	// gain/lose the changed peer. Unlike ChangedGroupIDs, a LinkGroup is not added to
+	// the output, so a one-sided membership change never wakes the whole group.
+	LinkGroups []string
 }
 
 func (c Change) isEmpty() bool {
@@ -186,7 +210,9 @@ func (c Change) isEmpty() bool {
 		len(c.Networks) == 0 &&
 		len(c.PostureCheckIDs) == 0 &&
 		len(c.DistributionGroupIDs) == 0 &&
-		len(c.RemovedPeersByGroup) == 0
+		len(c.RemovedPeersByGroup) == 0 &&
+		len(c.LinkGroups) == 0 &&
+		len(c.OutputPeerIDs) == 0
 }
 
 // Expand returns the deduplicated affected peer IDs from the preloaded Snapshot,
@@ -197,8 +223,8 @@ func (snap *Snapshot) Expand(ctx context.Context, accountID string, c Change) []
 		return nil
 	}
 	r := newResolver(ctx, snap, accountID, c)
-	log.WithContext(ctx).Tracef("affectedpeers expand start: account=%s changedGroups=%v changedPeers=%v policies=%d routes=%d routers=%d resources=%d networks=%d postureChecks=%v distributionGroups=%v",
+	log.WithContext(ctx).Tracef("affectedpeers expand start: account=%s changedGroups=%v changedPeers=%v linkGroups=%v policies=%d routes=%d routers=%d resources=%d networks=%d postureChecks=%v distributionGroups=%v",
-		accountID, c.ChangedGroupIDs, c.ChangedPeerIDs, len(c.Policies), len(c.Routes), len(c.Routers), len(c.Resources), len(c.Networks), c.PostureCheckIDs, c.DistributionGroupIDs)
+		accountID, c.ChangedGroupIDs, c.ChangedPeerIDs, c.LinkGroups, len(c.Policies), len(c.Routes), len(c.Routers), len(c.Resources), len(c.Networks), c.PostureCheckIDs, c.DistributionGroupIDs)
 	r.walk()
 	return r.expand()
 }
@@ -216,57 +242,84 @@ func Collect(ctx context.Context, s store.Store, accountID string, c Change) (gr
 	}
 	r := newResolver(ctx, snap, accountID, c)
 	r.walk()
-	return setToSlice(r.groupSet), setToSlice(r.peerSet)
+	return setToSlice(r.affectedGroups), setToSlice(r.affectedPeers)
 }
 
 func newResolver(ctx context.Context, snap *Snapshot, accountID string, c Change) *resolver {
 	r := &resolver{
-		ctx:             ctx,
+		ctx:            ctx,
-		snap:            snap,
+		snap:           snap,
-		accountID:       accountID,
+		accountID:      accountID,
-		change:          c,
+		change:         c,
-		changedGroupSet: toSet(c.ChangedGroupIDs),
+		linkGroups:     toSet(c.ChangedGroupIDs),
-		changedPeerSet:  toSet(c.ChangedPeerIDs),
+		outputGroups:   toSet(c.ChangedGroupIDs),
-		groupSet:        make(map[string]struct{}),
+		changedPeers:   toSet(c.ChangedPeerIDs),
-		peerSet:         make(map[string]struct{}),
+		affectedGroups: make(map[string]struct{}),
-		networkIDs:      make(map[string]struct{}),
+		affectedPeers:  make(map[string]struct{}),
 	}
+	// LinkGroups match policies/routes to find the opposite side but are NOT output:
+	// they go into linkGroups only, never outputGroups, so their members never fold in.
+	addAll(r.linkGroups, c.LinkGroups)
 	// Resolve each changed peer to its groups here so callers pass only ChangedPeerIDs.
 	r.seedChangedGroupsFromPeers()
-	r.matchedPolicies = append(r.matchedPolicies, c.Policies...)
 	return r
 }
 
-// seedChangedGroupsFromPeers adds each changed peer's groups to changedGroupSet so
+// seedChangedGroupsFromPeers adds each changed peer's groups to linkGroups so
 // the group-driven walkers fire for memberships, not just direct peer references.
+// These seeded groups are for MATCHING only — folding the changed entity's own
+// side is gated on outputGroups (the caller-reported groups), so a seeded group
+// never folds its whole membership; only the changed peer itself folds in.
 func (r *resolver) seedChangedGroupsFromPeers() {
-	if len(r.changedPeerSet) == 0 {
+	if len(r.changedPeers) == 0 {
 		return
 	}
 	for groupID, members := range r.snap.groupPeers {
-		for pID := range r.changedPeerSet {
+		for pID := range r.changedPeers {
 			if _, ok := members[pID]; ok {
-				r.changedGroupSet[groupID] = struct{}{}
+				r.linkGroups[groupID] = struct{}{}
 				break
 			}
 		}
 	}
 }
 
+// policySide selects which side of a policy rule to walk.
+type policySide int
+
+const (
+	sideSource policySide = iota
+	sideDestination
+)
+
+func (s policySide) opposite() policySide {
+	if s == sideSource {
+		return sideDestination
+	}
+	return sideSource
+}
+
+// walk resolves affected peers in two buckets, by how far each change propagates.
+//
+// BOTH-SIDES — the rule itself changed (an explicit policy edit, or a policy whose
+// posture check changed). Source AND destination refresh, so each such policy is
+// walked on both sides.
+//
+// OPPOSITE-SIDE — an endpoint moved but no rule changed. For each policy the change
+// touches we fold only the side AWAY from the change:
+//   - a changed peer/group sits ON a policy side -> fold the opposite side;
+//   - a changed router/resource/network sits on a NETWORK -> fold the SOURCE side of
+//     the policies whose destination reaches it (and the routers it implies).
+//
+// Routes, nameserver groups, DNS and embedded-proxy services distribute to their own
+// member peers, outside the policy graph, and are folded here too.
 func (r *resolver) walk() {
-	r.collectFromExplicitPolicies()
+	for _, policy := range r.bothSidesPolicies() {
-	r.collectFromExplicitRoutes(r.change.Routes)
+		r.foldPolicySide(policy, sideSource)
-	r.collectFromExplicitRouters(r.change.Routers)
+		r.foldPolicySide(policy, sideDestination)
-	r.collectFromExplicitResources(r.change.Resources)
+	}
-	r.collectFromExplicitNetworks(r.change.Networks)
-	r.collectFromPostureChecks(r.change.PostureCheckIDs)
 
-	// Distribution groups (nameserver/DNS) affect only their member peers: fold them
+	if len(r.linkGroups) > 0 || len(r.changedPeers) > 0 {
-	// straight into groupSet so expand() maps them to members, without the policy/
-	// route walk that changedGroupSet would trigger.
-	addAll(r.groupSet, r.change.DistributionGroupIDs)
-
-	if len(r.changedGroupSet) > 0 || len(r.changedPeerSet) > 0 {
 		r.collectFromPolicies()
 		r.collectFromRoutes()
 		r.collectFromNameServers()
@@ -275,7 +328,31 @@ func (r *resolver) walk() {
 		r.collectFromProxyServices()
 	}
 
-	r.collectResourceRouterBridge()
+	r.collectFromChangedRoutes(r.change.Routes)
+	r.collectFromChangedRouters(r.change.Routers)
+	r.collectFromChangedResources(r.change.Resources)
+	r.collectFromChangedNetworks(r.change.Networks)
+
+	// The explicitly changed peers always refresh their own maps. OnPeersUpdated only
+	// refreshes the resolver's output (it ignores the separately-passed changed peers),
+	// so the changed peer reaches its own new map only via here. An offline/deleted
+	// peer in the set is filtered downstream (filterConnectedAffectedPeers).
+	addAll(r.affectedPeers, setToSlice(r.changedPeers))
+	// OutputPeerIDs refresh themselves too, but unlike changedPeers their group
+	// memberships were not seeded into the walk (only the changed group was).
+	addAll(r.affectedPeers, r.change.OutputPeerIDs)
+
+	// Distribution groups (nameserver/DNS) affect only their member peers: fold them
+	// straight into affectedGroups so expand() maps them to members, without the
+	// policy/route walk that linkGroups would trigger.
+	addAll(r.affectedGroups, r.change.DistributionGroupIDs)
+}
+
+// bothSidesPolicies are the policies whose rule changed: the explicitly edited ones
+// plus those gated by a changed posture check. walk folds both their sides.
+func (r *resolver) bothSidesPolicies() []*types.Policy {
+	policies := append([]*types.Policy(nil), r.change.Policies...)
+	return r.appendPoliciesForPostureChecks(policies, r.change.PostureCheckIDs)
 }
 
 type resolver struct {
@@ -284,27 +361,71 @@ type resolver struct {
 	accountID string
 	change    Change
 
-	changedGroupSet map[string]struct{}
+	// Inputs — what changed. Set once at construction, read-only during the walk
-	changedPeerSet  map[string]struct{}
+	// (except linkGroups, which collectFromExplicitResources also seeds).
+	//
+	// linkGroups is the MATCH set: caller-changed groups ∪ the groups of changed
+	// peers ∪ changed-resource groups. A rule/route/router matches the change when
+	// one of its groups is here — used only to find the opposite side to fold.
+	//
+	// outputGroups is the FOLD-WHOLE-GROUP set: ONLY Change.ChangedGroupIDs. When a
+	// matched group is here, its whole membership is affected. A peer-seeded group
+	// is in linkGroups but NOT outputGroups, so it folds only the changed peer
+	// (changedPeers), never its siblings.
+	linkGroups   map[string]struct{}
+	outputGroups map[string]struct{}
+	changedPeers map[string]struct{}
 
-	groupSet map[string]struct{}
+	// Outputs — the answer. The only sets the walk accumulates into. affectedGroups
-	peerSet  map[string]struct{}
+	// is expanded to its member peers in expand().
-
+	affectedGroups map[string]struct{}
-	matchedPolicies []*types.Policy
+	affectedPeers  map[string]struct{}
-	networkIDs      map[string]struct{}
 }
 
-func (r *resolver) policies() []*types.Policy { return r.snap.policies }
+// policies returns the account's ENABLED policies from the snapshot. Disabled
+// policies grant no access, so the walk skips them when scanning existing account
+// data. Explicitly changed policies (Change.Policies, via bothSidesPolicies) are
+// processed regardless of Enabled, so disabling one still refreshes its peers.
+func (r *resolver) policies() []*types.Policy {
+	enabled := make([]*types.Policy, 0, len(r.snap.policies))
+	for _, policy := range r.snap.policies {
+		if policy != nil && policy.Enabled {
+			enabled = append(enabled, policy)
+		}
+	}
+	return enabled
+}
 
-func (r *resolver) networkResources() []*resourceTypes.NetworkResource { return r.snap.resources }
+// networkResources / networkRouters return the account's ENABLED resources/routers
+// from the snapshot. Disabled objects route to nobody, so the walk skips them when
+// it scans existing account data. The explicitly changed objects in the Change are
+// processed regardless of Enabled (collectFromChanged*), so disabling one still
+// refreshes the peers that lose access.
+func (r *resolver) networkResources() []*resourceTypes.NetworkResource {
+	enabled := make([]*resourceTypes.NetworkResource, 0, len(r.snap.resources))
+	for _, resource := range r.snap.resources {
+		if resource.Enabled {
+			enabled = append(enabled, resource)
+		}
+	}
+	return enabled
+}
 
-func (r *resolver) networkRouters() []*routerTypes.NetworkRouter { return r.snap.routers }
+func (r *resolver) networkRouters() []*routerTypes.NetworkRouter {
+	enabled := make([]*routerTypes.NetworkRouter, 0, len(r.snap.routers))
+	for _, router := range r.snap.routers {
+		if router.Enabled {
+			enabled = append(enabled, router)
+		}
+	}
+	return enabled
+}
 
 // peerIDsForGroups maps a group set to its member peer IDs via the preloaded index.
-func (r *resolver) peerIDsForGroups(groupSet map[string]struct{}) []string {
+func (r *resolver) peerIDsForGroups(groups map[string]struct{}) []string {
 	seen := make(map[string]struct{})
 	var ids []string
-	for gID := range groupSet {
+	for gID := range groups {
 		for pID := range r.snap.groupPeers[gID] {
 			if _, ok := seen[pID]; ok {
 				continue
@@ -317,25 +438,25 @@ func (r *resolver) peerIDsForGroups(groupSet map[string]struct{}) []string {
 }
 
 func (r *resolver) expand() []string {
-	peerIDs := r.peerIDsForGroups(r.groupSet)
+	peerIDs := r.peerIDsForGroups(r.affectedGroups)
 
 	log.WithContext(r.ctx).Tracef("affectedpeers expand: account=%s affectedGroups=%v -> %d group-member peers; direct peers=%v",
-		r.accountID, setToSlice(r.groupSet), len(peerIDs), setToSlice(r.peerSet))
+		r.accountID, setToSlice(r.affectedGroups), len(peerIDs), setToSlice(r.affectedPeers))
 
 	seen := make(map[string]struct{}, len(peerIDs))
 	for _, id := range peerIDs {
 		seen[id] = struct{}{}
 	}
-	for id := range r.peerSet {
+	for id := range r.affectedPeers {
 		if _, ok := seen[id]; !ok {
 			peerIDs = append(peerIDs, id)
 			seen[id] = struct{}{}
 		}
 	}
 
-	// Fold in removed peers only when their group is linked (in groupSet).
+	// Fold in removed peers only when their group is linked (in affectedGroups).
 	for groupID, removed := range r.change.RemovedPeersByGroup {
-		if _, linked := r.groupSet[groupID]; !linked {
+		if _, linked := r.affectedGroups[groupID]; !linked {
 			continue
 		}
 		for _, id := range removed {
@@ -351,169 +472,349 @@ func (r *resolver) expand() []string {
 	return peerIDs
 }
 
-func (r *resolver) collectFromExplicitPolicies() {
+// ruleSideGroups / ruleSideResource return the groups and the resource on the given
-	for _, policy := range r.matchedPolicies {
+// side of a rule.
-		if policy == nil {
+func ruleSideGroups(rule *types.PolicyRule, side policySide) []string {
-			continue
+	if side == sideDestination {
+		return rule.Destinations
+	}
+	return rule.Sources
+}
+
+func ruleSideResource(rule *types.PolicyRule, side policySide) types.Resource {
+	if side == sideDestination {
+		return rule.DestinationResource
+	}
+	return rule.SourceResource
+}
+
+// foldPolicySide folds one side of a policy down to affected peers: its groups
+// (resolved to members in expand) and its direct peer. When the side is the
+// DESTINATION and references a network resource (directly or via a destination
+// group's resources), it also folds the routers that serve that resource's network
+// — a destination resource is reached through its routers. A resource on the SOURCE
+// side routes to nobody (GetPoliciesForNetworkResource matches destinations only),
+// so the router hop is destination-only.
+func (r *resolver) foldPolicySide(policy *types.Policy, side policySide) {
+	if policy == nil {
+		return
+	}
+	for _, rule := range policy.Rules {
+		addAll(r.affectedGroups, ruleSideGroups(rule, side))
+		res := ruleSideResource(rule, side)
+		if res.Type == types.ResourceTypePeer && res.ID != "" {
+			r.affectedPeers[res.ID] = struct{}{}
 		}
-		log.WithContext(r.ctx).Tracef("collectFromExplicitPolicies: changed policy %s (%s) -> folding rule groups %v + direct peers",
+	}
-			policy.ID, policy.Name, policy.RuleGroups())
+	if side == sideDestination {
-		addAll(r.groupSet, policy.RuleGroups())
+		r.foldRoutersForResources(r.policyDestinationResourceIDs(policy))
-		collectPolicyDirectPeers(policy, r.peerSet)
 	}
 }
 
-func (r *resolver) collectFromExplicitRoutes(routes []*route.Route) {
+// appendPoliciesForPostureChecks appends every policy that references a changed
+// posture check (a rule change, so walk both sides).
+func (r *resolver) appendPoliciesForPostureChecks(policies []*types.Policy, postureCheckIDs []string) []*types.Policy {
+	if len(postureCheckIDs) == 0 {
+		return policies
+	}
+	ids := toSet(postureCheckIDs)
+	for _, policy := range r.policies() {
+		if !policyReferencesPostureChecks(policy, ids) || !policy.Enabled {
+			continue
+		}
+		log.WithContext(r.ctx).Tracef("appendPoliciesForPostureChecks: policy %s (%s) references changed posture checks %v -> both-sides policy",
+			policy.ID, policy.Name, postureCheckIDs)
+		policies = append(policies, policy)
+	}
+	return policies
+}
+
+// collectFromPolicies folds, for every policy whose rule a changed group or peer
+// touches, only the OPPOSITE side (down to peers, incl. destination routers), plus
+// the changed entity's own side: the changed group's whole membership when the
+// group itself changed (outputGroups), or the changed peer alone when matched via a
+// peer-seeded group (never its co-members).
+func (r *resolver) collectFromPolicies() {
+	for _, policy := range r.policies() {
+		for _, rule := range policy.Rules {
+			if !rule.Enabled {
+				continue // a disabled rule grants no access
+			}
+			r.foldRuleSideIfChanged(policy, rule, sideSource)
+			r.foldRuleSideIfChanged(policy, rule, sideDestination)
+		}
+	}
+}
+
+// foldRuleSideIfChanged: when a changed group or direct peer sits on `side` of the
+// rule, fold the opposite side fully (groups/peers + destination routers) and fold
+// the changed entity's own side (the whole changed group, or the changed peer alone).
+func (r *resolver) foldRuleSideIfChanged(policy *types.Policy, rule *types.PolicyRule, side policySide) {
+	nearGroups := ruleSideGroups(rule, side)
+	nearResource := ruleSideResource(rule, side)
+
+	matchedByGroup := anyInSet(nearGroups, r.linkGroups)
+	matchedByPeer := isDirectPeerInSet(nearResource, r.changedPeers)
+	if !matchedByGroup && !matchedByPeer {
+		return
+	}
+
+	// Opposite side, fully down to peers (a destination opposite also folds routers).
+	r.foldPolicySideForRule(policy, rule, side.opposite())
+
+	// Own side: fold the whole changed group's members only when the group itself
+	// changed (outputGroups). A peer-seeded or link-only group is not folded here —
+	// its siblings never refresh. The changed peers themselves are folded once, after
+	// the walk (see walk()).
+	for _, gID := range nearGroups {
+		if _, ok := r.outputGroups[gID]; ok {
+			r.affectedGroups[gID] = struct{}{}
+		}
+	}
+
+	// When the changed side IS a destination, the resources it targets are reached
+	// through their network's routers, so those routers refresh too (e.g. attaching a
+	// resource to a destination group, or a changed destination group/resource).
+	if side == sideDestination {
+		r.foldRoutersForResources(r.ruleDestinationResourceIDs(rule))
+	}
+}
+
+// foldPolicySideForRule folds one side of a single rule (groups + direct peer), and
+// for a destination side the routers of that rule's destination resources.
+func (r *resolver) foldPolicySideForRule(policy *types.Policy, rule *types.PolicyRule, side policySide) {
+	addAll(r.affectedGroups, ruleSideGroups(rule, side))
+	res := ruleSideResource(rule, side)
+	if res.Type == types.ResourceTypePeer && res.ID != "" {
+		r.affectedPeers[res.ID] = struct{}{}
+	}
+	if side == sideDestination {
+		r.foldRoutersForResources(r.ruleDestinationResourceIDs(rule))
+	}
+}
+
+// collectFromChangedRoutes folds an explicitly changed route's own groups and peer.
+func (r *resolver) collectFromChangedRoutes(routes []*route.Route) {
 	for _, rt := range routes {
 		if rt == nil {
 			continue
 		}
-		log.WithContext(r.ctx).Tracef("collectFromExplicitRoutes: changed route %s -> folding groups=%v peerGroups=%v accessControlGroups=%v peer=%q",
+		log.WithContext(r.ctx).Tracef("collectFromChangedRoutes: changed route %s -> folding groups=%v peerGroups=%v accessControlGroups=%v peer=%q",
 			rt.ID, rt.Groups, rt.PeerGroups, rt.AccessControlGroups, rt.Peer)
-		addAll(r.groupSet, rt.Groups, rt.PeerGroups, rt.AccessControlGroups)
+		addAll(r.affectedGroups, rt.Groups, rt.PeerGroups, rt.AccessControlGroups)
 		if rt.Peer != "" {
-			r.peerSet[rt.Peer] = struct{}{}
+			r.affectedPeers[rt.Peer] = struct{}{}
 		}
 	}
 }
 
-// collectFromExplicitRouters folds changed routers' peers and marks their networks
+// collectFromChangedRouters: a changed router refreshes its OWN backing peer/groups
-// for the bridge. Passing the old router keeps a repointed router's previous peers
+// (the changed entity) and the SOURCE side of every policy reaching a resource on
-// affected without a post-commit read.
+// its network (the router serves the whole network). Sibling routers on the network
-func (r *resolver) collectFromExplicitRouters(routers []*routerTypes.NetworkRouter) {
+// are independent and are NOT folded. Passing the old router state keeps a repointed
+// router's previous backing affected without a post-commit read.
+func (r *resolver) collectFromChangedRouters(routers []*routerTypes.NetworkRouter) {
 	for _, router := range routers {
 		if router == nil {
 			continue
 		}
-		log.WithContext(r.ctx).Tracef("collectFromExplicitRouters: changed router %s on network %s -> folding peerGroups=%v peer=%q and marking network for source bridge",
+		log.WithContext(r.ctx).Tracef("collectFromChangedRouters: changed router %s on network %s -> folding its own peerGroups=%v peer=%q + sources reaching network resources",
 			router.ID, router.NetworkID, router.PeerGroups, router.Peer)
-		addAll(r.groupSet, router.PeerGroups)
+		addAll(r.affectedGroups, router.PeerGroups)
 		if router.Peer != "" {
-			r.peerSet[router.Peer] = struct{}{}
+			r.affectedPeers[router.Peer] = struct{}{}
 		}
 		if router.NetworkID != "" {
-			r.networkIDs[router.NetworkID] = struct{}{}
+			r.foldPolicySourcesForResources(r.networkResourceIDs(router.NetworkID))
 		}
 	}
 }
 
-// collectFromExplicitResources marks changed resources' networks for the bridge and
+// collectFromChangedResources: a changed resource refreshes the SOURCE side of the
-// treats their group IDs as changed, so policies targeting the resource via a
+// policies targeting EXACTLY that resource — directly, or via one of the resource's
-// now-detached (old) group still refresh.
+// own groups (old∪new across the change, so a now-detached group's sources still
-func (r *resolver) collectFromExplicitResources(resources []*resourceTypes.NetworkResource) {
+// refresh) — plus the routers serving its network (the resource is reached through
+// them). It does not touch sibling resources on the same network.
+func (r *resolver) collectFromChangedResources(resources []*resourceTypes.NetworkResource) {
 	for _, resource := range resources {
 		if resource == nil {
 			continue
 		}
-		log.WithContext(r.ctx).Tracef("collectFromExplicitResources: changed resource %s on network %s -> marking network for bridge and treating groups %v as changed",
+		log.WithContext(r.ctx).Tracef("collectFromChangedResources: changed resource %s on network %s (groups %v) -> folding sources of policies targeting it + its network's routers",
 			resource.ID, resource.NetworkID, resource.GroupIDs)
-		addAll(r.changedGroupSet, resource.GroupIDs)
+		r.foldPolicySourcesForResource(resource.ID, resource.GroupIDs)
 		if resource.NetworkID != "" {
-			r.networkIDs[resource.NetworkID] = struct{}{}
+			r.foldRoutersOnNetworks(map[string]struct{}{resource.NetworkID: {}})
 		}
 	}
 }
 
-// collectFromExplicitNetworks marks changed networks for the bridge. A network has
+// foldPolicySourcesForResource folds the source side of every policy whose
-// no groups/peers of its own.
+// destination is the given resource — referenced directly, or via any of the given
-func (r *resolver) collectFromExplicitNetworks(networks []*networkTypes.Network) {
+// groups (the resource's own old∪new groups, which captures a detached group).
-	for _, network := range networks {
+func (r *resolver) foldPolicySourcesForResource(resourceID string, groupIDs []string) {
-		if network == nil {
+	groups := toSet(groupIDs)
+	for _, policy := range r.policies() {
+		if !policyTargetsResourceOrGroups(policy, resourceID, groups) {
 			continue
 		}
-		log.WithContext(r.ctx).Tracef("collectFromExplicitNetworks: changed network %s -> marking for bridge", network.ID)
+		log.WithContext(r.ctx).Tracef("foldPolicySourcesForResource: policy %s (%s) targets changed resource %s -> folding its source groups/peers", policy.ID, policy.Name, resourceID)
-		if network.ID != "" {
+		collectPolicySources(policy, r.affectedGroups, r.affectedPeers)
-			r.networkIDs[network.ID] = struct{}{}
-		}
 	}
 }
 
-func (r *resolver) collectFromPostureChecks(postureCheckIDs []string) {
+// policyTargetsResourceOrGroups reports whether a policy's destination is the given
-	if len(postureCheckIDs) == 0 {
+// resource directly, or one of the given destination groups.
+func policyTargetsResourceOrGroups(policy *types.Policy, resourceID string, groups map[string]struct{}) bool {
+	if policy == nil {
+		return false
+	}
+	for _, rule := range policy.Rules {
+		if !rule.Enabled {
+			continue
+		}
+		if rule.DestinationResource.Type != types.ResourceTypePeer && rule.DestinationResource.ID == resourceID && resourceID != "" {
+			return true
+		}
+		if anyInSet(rule.Destinations, groups) {
+			return true
+		}
+	}
+	return false
+}
+
+// collectFromChangedNetworks: a changed network refreshes the SOURCE side of the
+// policies reaching any of its resources, plus its routers. A network has no
+// groups/peers of its own.
+func (r *resolver) collectFromChangedNetworks(networks []*networkTypes.Network) {
+	for _, network := range networks {
+		if network == nil || network.ID == "" {
+			continue
+		}
+		log.WithContext(r.ctx).Tracef("collectFromChangedNetworks: changed network %s -> folding sources reaching its resources + its routers", network.ID)
+		resourceIDs := r.networkResourceIDs(network.ID)
+		r.foldPolicySourcesForResources(resourceIDs)
+		r.foldRoutersOnNetworks(map[string]struct{}{network.ID: {}})
+	}
+}
+
+// foldPolicySourcesForResources folds the source groups/peers of every policy whose
+// destination targets one of resourceIDs (directly or via a destination group).
+func (r *resolver) foldPolicySourcesForResources(resourceIDs map[string]struct{}) {
+	if len(resourceIDs) == 0 {
 		return
 	}
-	ids := toSet(postureCheckIDs)
 	for _, policy := range r.policies() {
-		if !policyReferencesPostureChecks(policy, ids) {
+		if r.policyTargetsResources(policy, resourceIDs) {
-			continue
+			log.WithContext(r.ctx).Tracef("foldPolicySourcesForResources: policy %s (%s) targets a changed resource -> folding its source groups/peers", policy.ID, policy.Name)
+			collectPolicySources(policy, r.affectedGroups, r.affectedPeers)
 		}
-		log.WithContext(r.ctx).Tracef("collectFromPostureChecks: policy %s (%s) references changed posture checks %v -> folding rule groups %v + direct peers",
-			policy.ID, policy.Name, postureCheckIDs, policy.RuleGroups())
-		addAll(r.groupSet, policy.RuleGroups())
-		collectPolicyDirectPeers(policy, r.peerSet)
-		r.matchedPolicies = append(r.matchedPolicies, policy)
-	}
-}
-
-func (r *resolver) collectFromPolicies() {
-	for _, policy := range r.policies() {
-		matchedByGroup := policyReferencesGroups(policy, r.changedGroupSet)
-		matchedByPeer := len(r.changedPeerSet) > 0 && policyReferencesDirectPeers(policy, r.changedPeerSet)
-		if !matchedByGroup && !matchedByPeer {
-			continue
-		}
-		log.WithContext(r.ctx).Tracef("collectFromPolicies: policy %s (%s) matched (byGroup=%t byPeer=%t) -> folding rule groups %v + direct peers",
-			policy.ID, policy.Name, matchedByGroup, matchedByPeer, policy.RuleGroups())
-		addAll(r.groupSet, policy.RuleGroups())
-		collectPolicyDirectPeers(policy, r.peerSet)
-		r.matchedPolicies = append(r.matchedPolicies, policy)
 	}
 }
 
+// collectFromRoutes folds, per matched route, the OPPOSITE side(s) fully and the
+// matched side's own groups only on a whole-group change (outputGroups). A route has
+// three peer sides — routing (Peer/PeerGroups), consumer (Groups) and ACL
+// (AccessControlGroups) — that each refresh the others; the changed side's own group
+// folds its siblings only when the group itself changed, never on a one-peer move.
 func (r *resolver) collectFromRoutes() {
 	for _, rt := range r.snap.routes {
-		matchedByGroup := anyInSet(rt.Groups, r.changedGroupSet) || anyInSet(rt.PeerGroups, r.changedGroupSet) || anyInSet(rt.AccessControlGroups, r.changedGroupSet)
+		if !rt.Enabled {
-		matchedByPeer := rt.Peer != "" && len(r.changedPeerSet) > 0 && isInSet(rt.Peer, r.changedPeerSet)
+			continue // disabled routes route to nobody; skip existing account data
-		if !matchedByGroup && !matchedByPeer {
+		}
+		routing := anyInSet(rt.PeerGroups, r.linkGroups) || (rt.Peer != "" && isInSet(rt.Peer, r.changedPeers))
+		consumer := anyInSet(rt.Groups, r.linkGroups)
+		acl := anyInSet(rt.AccessControlGroups, r.linkGroups)
+		if !routing && !consumer && !acl {
 			continue
 		}
-		log.WithContext(r.ctx).Tracef("collectFromRoutes: route %s matched (byGroup=%t byPeer=%t) -> folding groups=%v peerGroups=%v accessControlGroups=%v peer=%q",
+		log.WithContext(r.ctx).Tracef("collectFromRoutes: route %s matched (routing=%t consumer=%t acl=%t) -> folding opposite sides; own side gated on outputGroups",
-			rt.ID, matchedByGroup, matchedByPeer, rt.Groups, rt.PeerGroups, rt.AccessControlGroups, rt.Peer)
+			rt.ID, routing, consumer, acl)
-		addAll(r.groupSet, rt.Groups, rt.PeerGroups, rt.AccessControlGroups)
+		r.foldRouteSide(rt.PeerGroups, routing)
-		if rt.Peer != "" {
+		r.foldRouteSide(rt.Groups, consumer)
-			r.peerSet[rt.Peer] = struct{}{}
+		r.foldRouteSide(rt.AccessControlGroups, acl)
+		// The single routing Peer folds when the routing side is the OPPOSITE of the
+		// match (consumer/acl need it), or when that very peer is the change.
+		if rt.Peer != "" && (consumer || acl || isInSet(rt.Peer, r.changedPeers)) {
+			r.affectedPeers[rt.Peer] = struct{}{}
+		}
+	}
+}
+
+// foldRouteSide folds a route side: when this side is the one that matched, fold its
+// groups only on a whole-group change (outputGroups) so siblings of a single moved
+// peer stay put; otherwise it is an opposite side and folds fully.
+func (r *resolver) foldRouteSide(groups []string, matchedHere bool) {
+	if matchedHere {
+		r.foldOutputGroups(groups)
+		return
+	}
+	addAll(r.affectedGroups, groups)
+}
+
+// foldOutputGroups folds only the groups that the caller reported as wholly changed
+// (outputGroups). Used for a matched object's OWN side, where a peer-seeded or
+// link-only group must not pull in its siblings.
+func (r *resolver) foldOutputGroups(groups ...[]string) {
+	for _, gs := range groups {
+		for _, gID := range gs {
+			if _, ok := r.outputGroups[gID]; ok {
+				r.affectedGroups[gID] = struct{}{}
+			}
 		}
 	}
 }
 
 func (r *resolver) collectFromNameServers() {
-	if len(r.changedGroupSet) == 0 {
+	if len(r.linkGroups) == 0 {
 		return
 	}
 	for _, ns := range r.snap.nsGroups {
-		if anyInSet(ns.Groups, r.changedGroupSet) {
+		if anyInSet(ns.Groups, r.linkGroups) {
-			log.WithContext(r.ctx).Tracef("collectFromNameServers: nameserver group %s references a changed group -> folding its groups %v", ns.ID, ns.Groups)
+			// A nameserver group has no opposite side: a peer's DNS config depends only
-			addAll(r.groupSet, ns.Groups)
+			// on its own membership, so a one-peer move refreshes that peer alone (folded
+			// elsewhere). Fold the referenced groups only on a whole-group change.
+			log.WithContext(r.ctx).Tracef("collectFromNameServers: nameserver group %s references a linked group -> folding its groups %v (outputGroups only)", ns.ID, ns.Groups)
+			r.foldOutputGroups(ns.Groups)
 		}
 	}
 }
 
 func (r *resolver) collectFromDNSSettings() {
-	if len(r.changedGroupSet) == 0 || r.snap.dnsSettings == nil {
+	if len(r.linkGroups) == 0 || r.snap.dnsSettings == nil {
 		return
 	}
 	for _, gID := range r.snap.dnsSettings.DisabledManagementGroups {
-		if _, ok := r.changedGroupSet[gID]; ok {
+		if _, ok := r.linkGroups[gID]; ok {
 			log.WithContext(r.ctx).Tracef("collectFromDNSSettings: changed group %s is in DisabledManagementGroups -> folding it", gID)
-			r.groupSet[gID] = struct{}{}
+			r.affectedGroups[gID] = struct{}{}
 		}
 	}
 }
 
+// collectFromNetworkRouters handles a changed group/peer that BACKS a router (the
+// routing peer set moved): the router's own peers refresh and so do the sources of
+// the policies reaching its network's resources. Sibling routers on the network are
+// independent and are not folded.
 func (r *resolver) collectFromNetworkRouters() {
 	for _, router := range r.networkRouters() {
-		matchedByGroup := anyInSet(router.PeerGroups, r.changedGroupSet)
+		matchedByGroup := anyInSet(router.PeerGroups, r.linkGroups)
-		matchedByPeer := router.Peer != "" && len(r.changedPeerSet) > 0 && isInSet(router.Peer, r.changedPeerSet)
+		matchedByPeer := router.Peer != "" && len(r.changedPeers) > 0 && isInSet(router.Peer, r.changedPeers)
 		if !matchedByGroup && !matchedByPeer {
 			continue
 		}
-		log.WithContext(r.ctx).Tracef("collectFromNetworkRouters: router %s on network %s matched (byGroup=%t byPeer=%t) -> folding peerGroups=%v peer=%q and marking network for source bridge",
+		log.WithContext(r.ctx).Tracef("collectFromNetworkRouters: router %s on network %s matched (byGroup=%t byPeer=%t) -> folding its peerGroups=%v peer=%q (own groups on outputGroups) + sources reaching network resources",
 			router.ID, router.NetworkID, matchedByGroup, matchedByPeer, router.PeerGroups, router.Peer)
-		addAll(r.groupSet, router.PeerGroups)
+		// The backing PeerGroups are the matched (own) side: fold them only on a
+		// whole-group change so a one-peer move does not wake sibling backing peers. The
+		// opposite side (policy sources reaching the network) is folded below.
+		r.foldOutputGroups(router.PeerGroups)
 		if router.Peer != "" {
-			r.peerSet[router.Peer] = struct{}{}
+			r.affectedPeers[router.Peer] = struct{}{}
+		}
+		if router.NetworkID != "" {
+			r.foldPolicySourcesForResources(r.networkResourceIDs(router.NetworkID))
 		}
-		r.networkIDs[router.NetworkID] = struct{}{}
 	}
 }
 
@@ -526,42 +827,48 @@ func (r *resolver) collectFromProxyServices() {
 	expanded := r.expandChangedPeersWithGroups()
 
 	for _, svc := range services {
-		if svc == nil {
+		if svc == nil || !svc.Enabled {
-			continue
+			continue // a disabled service proxies nothing; skip existing account data
 		}
 		proxyPeers := proxyByCluster[svc.ProxyCluster]
 		if len(proxyPeers) == 0 {
 			continue
 		}
 		matchedByPeer := serviceMatchesChangedPeers(svc, proxyPeers, expanded)
-		matchedByAccessGroup := anyInSet(svc.AccessGroups, r.changedGroupSet)
+		matchedByAccessGroup := anyInSet(svc.AccessGroups, r.linkGroups)
 		if !matchedByPeer && !matchedByAccessGroup {
 			continue
 		}
-		log.WithContext(r.ctx).Tracef("collectFromProxyServices: service %s (cluster=%s) matched (byProxyOrTargetPeer=%t byAccessGroup=%t) -> folding %d proxy peers, peer targets and access groups %v",
+		log.WithContext(r.ctx).Tracef("collectFromProxyServices: service %s (cluster=%s) matched (byProxyOrTargetPeer=%t byAccessGroup=%t) -> folding %d proxy peers, peer targets; access groups %v on outputGroups only",
 			svc.ID, svc.ProxyCluster, matchedByPeer, matchedByAccessGroup, len(proxyPeers), svc.AccessGroups)
 		for _, pid := range proxyPeers {
-			r.peerSet[pid] = struct{}{}
+			r.affectedPeers[pid] = struct{}{}
 		}
 		for _, target := range svc.Targets {
+			if !target.Enabled {
+				continue // a disabled target forwards nothing
+			}
 			if target.TargetType == rpservice.TargetTypePeer && target.TargetId != "" {
-				r.peerSet[target.TargetId] = struct{}{}
+				r.affectedPeers[target.TargetId] = struct{}{}
 			}
 		}
-		addAll(r.groupSet, svc.AccessGroups)
+		// AccessGroups are the matched (own) side with no opposite to fold: a member's
+		// proxy access is self-contained, so a one-peer move refreshes that peer alone.
+		// Fold the groups only on a whole-group change.
+		r.foldOutputGroups(svc.AccessGroups)
 	}
 }
 
 func (r *resolver) expandChangedPeersWithGroups() map[string]struct{} {
-	if len(r.changedGroupSet) == 0 {
+	if len(r.linkGroups) == 0 {
-		return r.changedPeerSet
+		return r.changedPeers
 	}
-	ids := r.peerIDsForGroups(r.changedGroupSet)
+	ids := r.peerIDsForGroups(r.linkGroups)
 	if len(ids) == 0 {
-		return r.changedPeerSet
+		return r.changedPeers
 	}
-	merged := make(map[string]struct{}, len(r.changedPeerSet)+len(ids))
+	merged := make(map[string]struct{}, len(r.changedPeers)+len(ids))
-	for id := range r.changedPeerSet {
+	for id := range r.changedPeers {
 		merged[id] = struct{}{}
 	}
 	for _, id := range ids {
@@ -570,54 +877,36 @@ func (r *resolver) expandChangedPeersWithGroups() map[string]struct{} {
 	return merged
 }
 
-// collectResourceRouterBridge crosses between source peers and routing peers, which
+// foldRoutersForResources folds the routers serving the networks of the given
-// are reachable only via resource -> network -> router, not through the policy's own
+// resources (a destination resource is reached through its network's routers). It is
-// groups: source -> router (targeted resources' networks), then router -> source.
+// the resource -> network -> router hop used by foldPolicySide for a destination.
-func (r *resolver) collectResourceRouterBridge() {
+func (r *resolver) foldRoutersForResources(resourceIDs map[string]struct{}) {
-	r.bridgeSourceToRouters()
-	r.bridgeRoutersToSources()
-}
-
-func (r *resolver) bridgeSourceToRouters() {
-	resourceIDs := r.policyDestinationResourceIDs(r.matchedPolicies...)
 	if len(resourceIDs) == 0 {
 		return
 	}
-
+	r.foldRoutersOnNetworks(r.resourceNetworkIDs(resourceIDs))
-	networkIDs := r.resourceNetworkIDs(resourceIDs)
-	log.WithContext(r.ctx).Tracef("bridgeSourceToRouters: targeted resources %v -> networks %v (their routers become affected via the router->source pass)",
-		setToSlice(resourceIDs), setToSlice(networkIDs))
-	for id := range networkIDs {
-		r.networkIDs[id] = struct{}{}
-	}
 }
 
-func (r *resolver) bridgeRoutersToSources() {
+// ruleDestinationResourceIDs returns the destination resource IDs of a single rule:
-	if len(r.networkIDs) == 0 {
+// the direct DestinationResource plus the resources of its destination groups.
-		return
+func (r *resolver) ruleDestinationResourceIDs(rule *types.PolicyRule) map[string]struct{} {
+	resourceIDs := make(map[string]struct{})
+	if rule.DestinationResource.Type != types.ResourceTypePeer && rule.DestinationResource.ID != "" {
+		resourceIDs[rule.DestinationResource.ID] = struct{}{}
 	}
+	r.addGroupResourceIDs(toSet(rule.Destinations), resourceIDs)
+	return resourceIDs
+}
 
-	log.WithContext(r.ctx).Tracef("bridgeRoutersToSources: affected networks %v -> folding their routing peers and the source peers of policies targeting their resources",
+// networkResourceIDs returns the IDs of all resources on the given network.
-		setToSlice(r.networkIDs))
+func (r *resolver) networkResourceIDs(networkID string) map[string]struct{} {
-
-	r.foldRoutersOnNetworks(r.networkIDs)
-
 	resourceIDs := make(map[string]struct{})
 	for _, resource := range r.networkResources() {
-		if _, ok := r.networkIDs[resource.NetworkID]; ok {
+		if resource.NetworkID == networkID {
 			resourceIDs[resource.ID] = struct{}{}
 		}
 	}
-	if len(resourceIDs) == 0 {
+	return resourceIDs
-		return
-	}
-
-	for _, policy := range r.policies() {
-		if r.policyTargetsResources(policy, resourceIDs) {
-			log.WithContext(r.ctx).Tracef("bridgeRoutersToSources: policy %s (%s) targets an affected-network resource -> folding its source groups/peers", policy.ID, policy.Name)
-			collectPolicySources(policy, r.groupSet, r.peerSet)
-		}
-	}
 }
 
 func (r *resolver) foldRoutersOnNetworks(networkIDs map[string]struct{}) {
@@ -627,9 +916,9 @@ func (r *resolver) foldRoutersOnNetworks(networkIDs map[string]struct{}) {
 		}
 		log.WithContext(r.ctx).Tracef("bridgeRoutersToSources: router %s serves affected network %s -> folding peerGroups=%v peer=%q",
 			router.ID, router.NetworkID, router.PeerGroups, router.Peer)
-		addAll(r.groupSet, router.PeerGroups)
+		addAll(r.affectedGroups, router.PeerGroups)
 		if router.Peer != "" {
-			r.peerSet[router.Peer] = struct{}{}
+			r.affectedPeers[router.Peer] = struct{}{}
 		}
 	}
 }
@@ -650,6 +939,9 @@ func (r *resolver) policyTargetsResources(policy *types.Policy, resourceIDs map[
 	}
 	destGroupSet := make(map[string]struct{})
 	for _, rule := range policy.Rules {
+		if !rule.Enabled {
+			continue
+		}
 		if rule.DestinationResource.Type != types.ResourceTypePeer && isInSet(rule.DestinationResource.ID, resourceIDs) {
 			return true
 		}
@@ -714,44 +1006,20 @@ func (r *resolver) addGroupResourceIDs(groupIDs map[string]struct{}, resourceIDs
 	}
 }
 
-func collectPolicyDirectPeers(policy *types.Policy, peerSet map[string]struct{}) {
+// collectPolicySources folds the source groups/peers of a snapshot policy's enabled
+// rules (a disabled rule grants no access).
+func collectPolicySources(policy *types.Policy, groups, peers map[string]struct{}) {
 	for _, rule := range policy.Rules {
+		if !rule.Enabled {
+			continue
+		}
+		addAll(groups, rule.Sources)
 		if rule.SourceResource.Type == types.ResourceTypePeer && rule.SourceResource.ID != "" {
-			peerSet[rule.SourceResource.ID] = struct{}{}
+			peers[rule.SourceResource.ID] = struct{}{}
-		}
-		if rule.DestinationResource.Type == types.ResourceTypePeer && rule.DestinationResource.ID != "" {
-			peerSet[rule.DestinationResource.ID] = struct{}{}
 		}
 	}
 }
 
-func collectPolicySources(policy *types.Policy, groupSet, peerSet map[string]struct{}) {
-	for _, rule := range policy.Rules {
-		addAll(groupSet, rule.Sources)
-		if rule.SourceResource.Type == types.ResourceTypePeer && rule.SourceResource.ID != "" {
-			peerSet[rule.SourceResource.ID] = struct{}{}
-		}
-	}
-}
-
-func policyReferencesGroups(policy *types.Policy, groupSet map[string]struct{}) bool {
-	for _, rule := range policy.Rules {
-		if anyInSet(rule.Sources, groupSet) || anyInSet(rule.Destinations, groupSet) {
-			return true
-		}
-	}
-	return false
-}
-
-func policyReferencesDirectPeers(policy *types.Policy, changedSet map[string]struct{}) bool {
-	for _, rule := range policy.Rules {
-		if isDirectPeerInSet(rule.SourceResource, changedSet) || isDirectPeerInSet(rule.DestinationResource, changedSet) {
-			return true
-		}
-	}
-	return false
-}
-
 func policyReferencesPostureChecks(policy *types.Policy, ids map[string]struct{}) bool {
 	for _, id := range policy.SourcePostureChecks {
 		if _, ok := ids[id]; ok {
@@ -776,7 +1044,7 @@ func serviceMatchesChangedPeers(svc *rpservice.Service, proxyPeers []string, cha
 		}
 	}
 	for _, target := range svc.Targets {
-		if target.TargetType != rpservice.TargetTypePeer || target.TargetId == "" {
+		if !target.Enabled || target.TargetType != rpservice.TargetTypePeer || target.TargetId == "" {
 			continue
 		}
 		if _, ok := changedPeers[target.TargetId]; ok {

management/server/affectedpeers/resolver_test.go

@@ -10,8 +10,8 @@ import (
 	"github.com/netbirdio/netbird/management/server/types"
 )
 
-// policyGroupsAndPeers mirrors the explicit-policy extraction (RuleGroups +
+// policyGroupsAndPeers mirrors the both-sides extraction (RuleGroups + direct peers)
-// direct peers) the resolver folds in, for asserting the pure logic.
+// the resolver folds in for a changed policy, for asserting the pure logic.
 func policyGroupsAndPeers(policies ...*types.Policy) (groups []string, peers []string) {
 	peerSet := map[string]struct{}{}
 	for _, p := range policies {
@@ -19,7 +19,14 @@ func policyGroupsAndPeers(policies ...*types.Policy) (groups []string, peers []s
 			continue
 		}
 		groups = append(groups, p.RuleGroups()...)
-		collectPolicyDirectPeers(p, peerSet)
+		for _, rule := range p.Rules {
+			if rule.SourceResource.Type == types.ResourceTypePeer && rule.SourceResource.ID != "" {
+				peerSet[rule.SourceResource.ID] = struct{}{}
+			}
+			if rule.DestinationResource.Type == types.ResourceTypePeer && rule.DestinationResource.ID != "" {
+				peerSet[rule.DestinationResource.ID] = struct{}{}
+			}
+		}
 	}
 	for id := range peerSet {
 		peers = append(peers, id)
@@ -80,26 +87,6 @@ func TestChangeIsEmpty(t *testing.T) {
 	assert.False(t, Change{PostureCheckIDs: []string{"pc"}}.isEmpty())
 }
 
-func TestPolicyReferencesGroups(t *testing.T) {
-	policy := &types.Policy{Rules: []*types.PolicyRule{{Sources: []string{"g1", "g2"}, Destinations: []string{"g3"}}}}
-
-	assert.True(t, policyReferencesGroups(policy, map[string]struct{}{"g1": {}}))
-	assert.True(t, policyReferencesGroups(policy, map[string]struct{}{"g3": {}}))
-	assert.False(t, policyReferencesGroups(policy, map[string]struct{}{"g4": {}}))
-	assert.False(t, policyReferencesGroups(policy, map[string]struct{}{}))
-}
-
-func TestPolicyReferencesDirectPeers(t *testing.T) {
-	policy := &types.Policy{Rules: []*types.PolicyRule{{
-		SourceResource:      types.Resource{Type: types.ResourceTypePeer, ID: "p1"},
-		DestinationResource: types.Resource{Type: types.ResourceTypeHost, ID: "r1"},
-	}}}
-
-	assert.True(t, policyReferencesDirectPeers(policy, map[string]struct{}{"p1": {}}))
-	assert.False(t, policyReferencesDirectPeers(policy, map[string]struct{}{"r1": {}}))
-	assert.False(t, policyReferencesDirectPeers(policy, map[string]struct{}{"p2": {}}))
-}
-
 func TestPolicyReferencesPostureChecks(t *testing.T) {
 	policy := &types.Policy{SourcePostureChecks: []string{"pc1", "pc2"}}
 
@@ -107,24 +94,9 @@ func TestPolicyReferencesPostureChecks(t *testing.T) {
 	assert.False(t, policyReferencesPostureChecks(policy, map[string]struct{}{"pc3": {}}))
 }
 
-func TestCollectPolicyDirectPeers(t *testing.T) {
-	policy := &types.Policy{Rules: []*types.PolicyRule{{
-		SourceResource:      types.Resource{Type: types.ResourceTypePeer, ID: "p1"},
-		DestinationResource: types.Resource{Type: types.ResourceTypePeer, ID: "p2"},
-	}, {
-		DestinationResource: types.Resource{Type: types.ResourceTypeHost, ID: "r1"},
-	}}}
-
-	peerSet := map[string]struct{}{}
-	collectPolicyDirectPeers(policy, peerSet)
-
-	assert.Contains(t, peerSet, "p1")
-	assert.Contains(t, peerSet, "p2")
-	assert.NotContains(t, peerSet, "r1")
-}
-
 func TestCollectPolicySources(t *testing.T) {
 	policy := &types.Policy{Rules: []*types.PolicyRule{{
+		Enabled:        true,
 		Sources:        []string{"g1"},
 		SourceResource: types.Resource{Type: types.ResourceTypePeer, ID: "p1"},
 		Destinations:   []string{"g2"},

management/server/group.go

@@ -520,7 +520,12 @@ func collectDeletableGroups(ctx context.Context, transaction store.Store, accoun
 // GroupAddPeer appends peer to the group
 func (am *DefaultAccountManager) GroupAddPeer(ctx context.Context, accountID, groupID, peerID string) error {
 	var snap *affectedpeers.Snapshot
-	change := affectedpeers.Change{ChangedGroupIDs: []string{groupID}}
+	// A membership change affects only the peer itself and the opposite side of THIS
+	// group's policies — not the group's other members, and not the peer's other
+	// groups. LinkGroups walks only this group (matched, not expanded); OutputPeerIDs
+	// refreshes the peer without seeding its other group memberships. For an
+	// intra-group policy the opposite side is the group, so its members still refresh.
+	change := affectedpeers.Change{OutputPeerIDs: []string{peerID}, LinkGroups: []string{groupID}}
 
 	err := am.Store.ExecuteInTransaction(ctx, func(transaction store.Store) error {
 		if err := transaction.AddPeerToGroup(ctx, accountID, peerID, groupID); err != nil {
@@ -586,10 +591,11 @@ func (am *DefaultAccountManager) GroupAddResource(ctx context.Context, accountID
 // GroupDeletePeer removes peer from the group
 func (am *DefaultAccountManager) GroupDeletePeer(ctx context.Context, accountID, groupID, peerID string) error {
 	var snap *affectedpeers.Snapshot
-	change := affectedpeers.Change{
+	// Same as GroupAddPeer: the removed peer and the opposite side of THIS group's
-		ChangedGroupIDs:     []string{groupID},
+	// policies refresh, not the group's other members or the peer's other groups. The
-		RemovedPeersByGroup: map[string][]string{groupID: {peerID}},
+	// peer is no longer in the group's index, but LinkGroups still drives the
-	}
+	// opposite-side walk, and OutputPeerIDs refreshes the removed peer itself.
+	change := affectedpeers.Change{OutputPeerIDs: []string{peerID}, LinkGroups: []string{groupID}}
 
 	err := am.Store.ExecuteInTransaction(ctx, func(transaction store.Store) error {
 		if err := transaction.RemovePeerFromGroup(ctx, peerID, groupID); err != nil {
@@ -600,8 +606,6 @@ func (am *DefaultAccountManager) GroupDeletePeer(ctx context.Context, accountID,
 			return err
 		}
 
-		// The removed peer is carried in change.RemovedPeersByGroup and folded in
-		// only when the group is linked, so loading post-removal is correct.
 		var err error
 		if snap, err = affectedpeers.Load(ctx, transaction, accountID, change); err != nil {
 			return err

management/server/http/handlers/users/users_handler.go

@@ -220,7 +220,7 @@ func (h *handler) getAllUsers(w http.ResponseWriter, r *http.Request) {
 		}
 
 		includeServiceUser, err := strconv.ParseBool(serviceUser)
-		log.WithContext(r.Context()).Debugf("Should include service user: %v", includeServiceUser)
+		log.WithContext(r.Context()).Tracef("Should include service user: %v", includeServiceUser)
 		if err != nil {
 			util.WriteError(r.Context(), status.Errorf(status.InvalidArgument, "invalid service_user query parameter"), w)
 			return

management/server/peer.go

@@ -209,14 +209,14 @@ func (am *DefaultAccountManager) resolvePeerLocation(ctx context.Context, peer *
 	if am.geo == nil || realIP == nil {
 		return nil
 	}
-	if peer.Location.ConnectionIP != nil && peer.Location.ConnectionIP.Equal(realIP) {
-		return nil
-	}
 	location, err := am.geo.Lookup(realIP)
 	if err != nil {
 		log.WithContext(ctx).Warnf("failed to get location for peer %s realip: [%s]: %v", peer.ID, realIP.String(), err)
 		return nil
 	}
+	if peer.Location.ConnectionIP != nil && peer.Location.ConnectionIP.Equal(realIP) && peer.Location.GeoNameID == location.City.GeonameID {
+		return nil
+	}
 	return &nbpeer.Location{
 		ConnectionIP: realIP,
 		CountryCode:  location.Country.ISOCode,
@@ -730,7 +730,7 @@ func (am *DefaultAccountManager) handleSetupKeyAddedPeer(ctx context.Context, en
 func (am *DefaultAccountManager) AddPeer(ctx context.Context, accountID, setupKey, userID string, peer *nbpeer.Peer, temporary bool) (*nbpeer.Peer, *types.Network, []*posture.Checks, bool, error) {
 	if setupKey == "" && userID == "" && !peer.ProxyMeta.Embedded {
 		// no auth method provided => reject access
-		return nil, nil, nil, false, status.Errorf(status.Unauthenticated, "no peer auth method provided, please use a setup key or interactive SSO login")
+		return nil, nil, nil, false, status.ErrNoAuthMethodProvided
 	}
 
 	upperKey := strings.ToUpper(setupKey)
@@ -1051,8 +1051,8 @@ func (am *DefaultAccountManager) SyncPeer(ctx context.Context, sync types.PeerSy
 		return nil, nil, nil, 0, err
 	}
 
-	metaDiffAffectsPosture := posture.AffectsPosture(&metaDiff, resPostureChecks)
+	metaDiffAffectsPosture := posture.AffectsPosture(ctx, &metaDiff, resPostureChecks)
-	if isStatusChanged || sync.UpdateAccountPeers || ipv6CapabilityChanged || metaDiffAffectsPosture || metaDiff.VersionChanged || metaDiff.Hostname {
+	if requiresPeerUpdate(ctx, isStatusChanged, sync.UpdateAccountPeers, ipv6CapabilityChanged, metaDiffAffectsPosture, metaDiff.VersionChanged(), metaDiff.HostnameChanged()) {
 		changedPeerIDs := []string{peer.ID}
 		affectedPeerIDs := am.syncPeerAffectedPeers(ctx, accountID, peer.ID, nmap, peerNotValid, metaDiffAffectsPosture)
 		if err = am.networkMapController.OnPeersUpdated(ctx, accountID, changedPeerIDs, affectedPeerIDs); err != nil {
@@ -1063,6 +1063,29 @@ func (am *DefaultAccountManager) SyncPeer(ctx context.Context, sync types.PeerSy
 	return peer, nmap, resPostureChecks, dnsFwdPort, nil
 }
 
+func requiresPeerUpdate(ctx context.Context, isStatusChanged, updateAccountPeers, ipv6CapabilityChanged, metaDiffAffectsPosture, versionChanged, hostname bool) bool {
+	var reason string
+	switch {
+	case isStatusChanged:
+		reason = "status changed"
+	case updateAccountPeers:
+		reason = "update account peers"
+	case ipv6CapabilityChanged:
+		reason = "ipv6 capability changed"
+	case metaDiffAffectsPosture:
+		reason = "meta diff affects posture"
+	case versionChanged:
+		reason = "version changed"
+	case hostname:
+		reason = "hostname changed"
+	default:
+		return false
+	}
+
+	log.WithContext(ctx).Tracef("peer update required: %s", reason)
+	return true
+}
+
 // syncPeerAffectedPeers resolves the peers affected by a SyncPeer change. The
 // peer's own validated network map is bidirectional for policy and routing
 // reachability, so when the peer stays valid and no source-posture gate is in

management/server/peer/peer.go

@@ -107,6 +107,15 @@ type Location struct {
 	GeoNameID    uint // city level geoname id
 }
 
+// equal reports whether two locations match. ConnectionIP is a net.IP slice, so it uses
+// IP.Equal, not ==.
+func (l Location) equal(other Location) bool {
+	return l.CountryCode == other.CountryCode &&
+		l.CityName == other.CityName &&
+		l.GeoNameID == other.GeoNameID &&
+		l.ConnectionIP.Equal(other.ConnectionIP)
+}
+
 // NetworkAddress is the IP address with network and MAC address of a network interface
 type NetworkAddress struct {
 	NetIP netip.Prefix `gorm:"serializer:json"`
@@ -267,185 +276,141 @@ func (p *Peer) UpdateMetaIfNew(ctx context.Context, meta PeerSystemMeta, newLoca
 		return MetaDiff{}
 	}
 
-	versionChanged := p.Meta.WtVersion != meta.WtVersion
-
 	// Avoid overwriting UIVersion if the update was triggered sole by the CLI client
 	if meta.UIVersion == "" {
 		meta.UIVersion = p.Meta.UIVersion
 	}
 
-	oldVersion := p.Meta.WtVersion
+	effectiveLocation := p.Location
+	if newLocation != nil {
+		effectiveLocation = *newLocation
+	}
 
-	diff := diffMeta(p.Meta, meta)
+	diff := diffMeta(p.Meta, meta, p.Location, effectiveLocation)
-	if diff.Any() {
+	if diff.Updated() {
 		p.Meta = meta
 	}
-	diff.VersionChanged = versionChanged
+	p.Location = effectiveLocation
 
-	locationInfo := ""
+	if diff.Updated() {
-	if newLocation != nil {
+		log.WithContext(ctx).Debug(diff.LogSummary())
-		p.Location = *newLocation
-		diff.LocationChanged = true
-		locationInfo = fmt.Sprintf("location changed to %s, ", newLocation.ConnectionIP)
-	}
-
-	versionInfo := ""
-	if diff.VersionChanged {
-		versionInfo = fmt.Sprintf("version changed: %s -> %s, ", oldVersion, meta.WtVersion)
-	}
-
-	if diff.Any() || diff.VersionChanged || diff.LocationChanged {
-		log.WithContext(ctx).
-			Debugf("peer meta updated, %s%s%d field(s) changed: %s", versionInfo, locationInfo, len(diff.Changed), strings.Join(diff.Changed, ", "))
 	}
 
 	return diff
 }
 
-// MetaDiff records which PeerSystemMeta fields differ between two metas. Each bool
+// MetaDiff holds a peer's full before/after state across a sync: both metas and both
-// maps to a single struct field, except Environment, which is split into Cloud and
+// connection locations (the location lives on Peer, not PeerSystemMeta, but posture
-// Platform. Changed holds the human-readable `field: <old> -> <new>` entries so the
+// checks read it). Changed lists what moved, for logging and the persistence decision;
-// existing log line and isEqual can be derived from the same comparison.
+// the snapshots let a posture check be replayed against old and new. Everything is derived
-//
+// from these fields, so there are no parallel per-field flags to keep in sync.
-// VersionChanged and LocationChanged sit outside the per-meta-field set:
-// VersionChanged tracks the WireGuard client version specifically (compared before
-// the UIVersion fixup, to signal client upgrades) and LocationChanged tracks the
-// peer's connection geo location, which lives on Peer rather than PeerSystemMeta.
-// Neither contributes an entry to Changed, so the field-coverage accounting stays
-// driven purely by the PeerSystemMeta comparison.
 type MetaDiff struct {
-	Hostname            bool
+	OldMeta     PeerSystemMeta
-	GoOS                bool
+	NewMeta     PeerSystemMeta
-	Kernel              bool
+	OldLocation Location
-	KernelVersion       bool
+	NewLocation Location
-	Core                bool
-	Platform            bool
-	OS                  bool
-	OSVersion           bool
-	WtVersion           bool
-	UIVersion           bool
-	SystemSerialNumber  bool
-	SystemProductName   bool
-	SystemManufacturer  bool
-	EnvironmentCloud    bool
-	EnvironmentPlatform bool
-	Flags               bool
-	Capabilities        bool
-	NetworkAddresses    bool
-	Files               bool
-
-	VersionChanged  bool
-	LocationChanged bool
 
 	Changed []string
 }
 
-// Any reports whether any PeerSystemMeta field changed.
+// Updated reports whether anything changed and the peer must be persisted. diffMeta fills
-func (d MetaDiff) Any() bool {
+// Changed in the pass that builds the diff, so this is a length check, not a re-comparison.
+// Pointer receiver: MetaDiff embeds two metas, so copying it per call is wasteful.
+func (d *MetaDiff) Updated() bool {
 	return len(d.Changed) != 0
 }
 
-// Updated reports whether the peer needs to be persisted: any meta field changed
+// VersionChanged reports whether the WireGuard client version changed (a client upgrade).
-// or the geo location changed. The version flag alone does not imply a write,
+func (d *MetaDiff) VersionChanged() bool {
-// since a version change is also reflected in the WtVersion meta field.
+	return d.OldMeta.WtVersion != d.NewMeta.WtVersion
-func (d MetaDiff) Updated() bool {
+}
-	return d.Any() || d.LocationChanged || d.VersionChanged
+
+// HostnameChanged reports whether the peer's hostname changed.
+func (d *MetaDiff) HostnameChanged() bool {
+	return d.OldMeta.Hostname != d.NewMeta.Hostname
+}
+
+// LogSummary renders the changed fields as a single human-readable line.
+func (d *MetaDiff) LogSummary() string {
+	return fmt.Sprintf("peer meta updated, %d field(s) changed: %s",
+		len(d.Changed), strings.Join(d.Changed, ", "))
 }
 
 func metaDiff(oldMeta, newMeta PeerSystemMeta) []string {
-	return diffMeta(oldMeta, newMeta).Changed
+	return diffMeta(oldMeta, newMeta, Location{}, Location{}).Changed
 }
 
-// diffMeta compares two metas field by field, returning both a per-field flag set
+// diffMeta snapshots a peer's old and new state and records a Changed entry per field that
-// (for callers that need to know exactly what changed, e.g. matching against
+// moved. It is the single source of truth for the comparison: isEqual is an empty Changed
-// posture checks) and the human-readable Changed list. It is the single source of
+// list, so the log line and the persistence decision can never disagree.
-// truth for meta comparison: isEqual reports equality as an empty diff, so the log
+func diffMeta(oldMeta, newMeta PeerSystemMeta, oldLocation, newLocation Location) MetaDiff {
-// line, the change decision, and the flags can never disagree.
+	d := MetaDiff{OldMeta: oldMeta, NewMeta: newMeta, OldLocation: oldLocation, NewLocation: newLocation}
-func diffMeta(oldMeta, newMeta PeerSystemMeta) MetaDiff {
-	var d MetaDiff
 	add := func(field string, oldVal, newVal any) {
 		d.Changed = append(d.Changed, fmt.Sprintf("%s: %v -> %v", field, oldVal, newVal))
 	}
 
 	if oldMeta.Hostname != newMeta.Hostname {
-		d.Hostname = true
 		add("hostname", oldMeta.Hostname, newMeta.Hostname)
 	}
 	if oldMeta.GoOS != newMeta.GoOS {
-		d.GoOS = true
 		add("goos", oldMeta.GoOS, newMeta.GoOS)
 	}
 	if oldMeta.Kernel != newMeta.Kernel {
-		d.Kernel = true
 		add("kernel", oldMeta.Kernel, newMeta.Kernel)
 	}
 	if oldMeta.KernelVersion != newMeta.KernelVersion {
-		d.KernelVersion = true
 		add("kernel_version", oldMeta.KernelVersion, newMeta.KernelVersion)
 	}
 	if oldMeta.Core != newMeta.Core {
-		d.Core = true
 		add("core", oldMeta.Core, newMeta.Core)
 	}
 	if oldMeta.Platform != newMeta.Platform {
-		d.Platform = true
 		add("platform", oldMeta.Platform, newMeta.Platform)
 	}
 	if oldMeta.OS != newMeta.OS {
-		d.OS = true
 		add("os", oldMeta.OS, newMeta.OS)
 	}
 	if oldMeta.OSVersion != newMeta.OSVersion {
-		d.OSVersion = true
 		add("os_version", oldMeta.OSVersion, newMeta.OSVersion)
 	}
 	if oldMeta.WtVersion != newMeta.WtVersion {
-		d.WtVersion = true
 		add("wt_version", oldMeta.WtVersion, newMeta.WtVersion)
 	}
 	if oldMeta.UIVersion != newMeta.UIVersion {
-		d.UIVersion = true
 		add("ui_version", oldMeta.UIVersion, newMeta.UIVersion)
 	}
 	if oldMeta.SystemSerialNumber != newMeta.SystemSerialNumber {
-		d.SystemSerialNumber = true
 		add("system_serial_number", oldMeta.SystemSerialNumber, newMeta.SystemSerialNumber)
 	}
 	if oldMeta.SystemProductName != newMeta.SystemProductName {
-		d.SystemProductName = true
 		add("system_product_name", oldMeta.SystemProductName, newMeta.SystemProductName)
 	}
 	if oldMeta.SystemManufacturer != newMeta.SystemManufacturer {
-		d.SystemManufacturer = true
 		add("system_manufacturer", oldMeta.SystemManufacturer, newMeta.SystemManufacturer)
 	}
 	if oldMeta.Environment.Cloud != newMeta.Environment.Cloud {
-		d.EnvironmentCloud = true
 		add("environment_cloud", oldMeta.Environment.Cloud, newMeta.Environment.Cloud)
 	}
 	if oldMeta.Environment.Platform != newMeta.Environment.Platform {
-		d.EnvironmentPlatform = true
 		add("environment_platform", oldMeta.Environment.Platform, newMeta.Environment.Platform)
 	}
 	if !oldMeta.Flags.isEqual(newMeta.Flags) {
-		d.Flags = true
 		add("flags", fmt.Sprintf("%+v", oldMeta.Flags), fmt.Sprintf("%+v", newMeta.Flags))
 	}
 	if !capabilitiesEqual(oldMeta.Capabilities, newMeta.Capabilities) {
-		d.Capabilities = true
 		add("capabilities", oldMeta.Capabilities, newMeta.Capabilities)
 	}
-
 	if !sameMultiset(oldMeta.NetworkAddresses, newMeta.NetworkAddresses) {
-		d.NetworkAddresses = true
 		add("network_addresses", fmt.Sprintf("%v", oldMeta.NetworkAddresses), fmt.Sprintf("%v", newMeta.NetworkAddresses))
 	}
-
 	if !sameMultiset(oldMeta.Files, newMeta.Files) {
-		d.Files = true
 		add("files", fmt.Sprintf("%v", oldMeta.Files), fmt.Sprintf("%v", newMeta.Files))
 	}
 
+	if !oldLocation.equal(newLocation) {
+		add("connection_ip", oldLocation.ConnectionIP, newLocation.ConnectionIP)
+	}
+
 	return d
 }
 

management/server/peer_test.go

@@ -49,6 +49,7 @@ import (
 
 	nbdns "github.com/netbirdio/netbird/dns"
 	"github.com/netbirdio/netbird/management/server/activity"
+	"github.com/netbirdio/netbird/management/server/geolocation"
 	nbpeer "github.com/netbirdio/netbird/management/server/peer"
 	"github.com/netbirdio/netbird/management/server/posture"
 	"github.com/netbirdio/netbird/management/server/store"
@@ -2893,3 +2894,141 @@ func TestUpdatePeer_DnsLabelUniqueName(t *testing.T) {
 	require.NoError(t, err, "renaming to unique FQDN should succeed")
 	assert.Equal(t, "api-server", updated.DNSLabel, "DNS label should be first label of FQDN")
 }
+
+// fakeGeo is a configurable geolocation.Geolocation implementation for tests. It
+// returns a record built from the configured city geoname id, or an error when set.
+type fakeGeo struct {
+	geoNameID uint
+	isoCode   string
+	cityName  string
+	err       error
+}
+
+func (g *fakeGeo) Lookup(net.IP) (*geolocation.Record, error) {
+	if g.err != nil {
+		return nil, g.err
+	}
+	record := &geolocation.Record{}
+	record.City.GeonameID = g.geoNameID
+	record.City.Names.En = g.cityName
+	record.Country.ISOCode = g.isoCode
+	return record, nil
+}
+
+func (g *fakeGeo) GetAllCountries() ([]geolocation.Country, error) { return nil, nil }
+
+func (g *fakeGeo) GetCitiesByCountry(string) ([]geolocation.City, error) { return nil, nil }
+
+func (g *fakeGeo) Stop() error { return nil }
+
+func TestResolvePeerLocation(t *testing.T) {
+	realIP := net.ParseIP("203.0.113.10")
+
+	tests := []struct {
+		name    string
+		geo     geolocation.Geolocation
+		peer    *nbpeer.Peer
+		realIP  net.IP
+		want    *nbpeer.Location
+		wantNil bool
+	}{
+		{
+			name:    "no geo configured returns nil",
+			geo:     nil,
+			peer:    &nbpeer.Peer{ID: "p1"},
+			realIP:  realIP,
+			wantNil: true,
+		},
+		{
+			name:    "nil real IP returns nil",
+			geo:     &fakeGeo{geoNameID: 100},
+			peer:    &nbpeer.Peer{ID: "p1"},
+			realIP:  nil,
+			wantNil: true,
+		},
+		{
+			name:    "lookup error returns nil",
+			geo:     &fakeGeo{err: fmt.Errorf("lookup boom")},
+			peer:    &nbpeer.Peer{ID: "p1"},
+			realIP:  realIP,
+			wantNil: true,
+		},
+		{
+			name: "same IP and same geoname returns nil",
+			geo:  &fakeGeo{geoNameID: 100, isoCode: "US", cityName: "City A"},
+			peer: &nbpeer.Peer{
+				ID: "p1",
+				Location: nbpeer.Location{
+					ConnectionIP: realIP,
+					GeoNameID:    100,
+				},
+			},
+			realIP:  realIP,
+			wantNil: true,
+		},
+		{
+			name: "same IP but changed geoname returns location",
+			geo:  &fakeGeo{geoNameID: 200, isoCode: "US", cityName: "City B"},
+			peer: &nbpeer.Peer{
+				ID: "p1",
+				Location: nbpeer.Location{
+					ConnectionIP: realIP,
+					GeoNameID:    100,
+				},
+			},
+			realIP: realIP,
+			want: &nbpeer.Location{
+				ConnectionIP: realIP,
+				CountryCode:  "US",
+				CityName:     "City B",
+				GeoNameID:    200,
+			},
+		},
+		{
+			name: "different IP returns location",
+			geo:  &fakeGeo{geoNameID: 100, isoCode: "US", cityName: "City A"},
+			peer: &nbpeer.Peer{
+				ID: "p1",
+				Location: nbpeer.Location{
+					ConnectionIP: net.ParseIP("198.51.100.7"),
+					GeoNameID:    100,
+				},
+			},
+			realIP: realIP,
+			want: &nbpeer.Location{
+				ConnectionIP: realIP,
+				CountryCode:  "US",
+				CityName:     "City A",
+				GeoNameID:    100,
+			},
+		},
+		{
+			name:   "no prior location returns location",
+			geo:    &fakeGeo{geoNameID: 100, isoCode: "US", cityName: "City A"},
+			peer:   &nbpeer.Peer{ID: "p1"},
+			realIP: realIP,
+			want: &nbpeer.Location{
+				ConnectionIP: realIP,
+				CountryCode:  "US",
+				CityName:     "City A",
+				GeoNameID:    100,
+			},
+		},
+	}
+
+	for _, tt := range tests {
+		t.Run(tt.name, func(t *testing.T) {
+			am := &DefaultAccountManager{geo: tt.geo}
+			got := am.resolvePeerLocation(context.Background(), tt.peer, tt.realIP)
+			if tt.wantNil {
+				assert.Nil(t, got, "resolved location should be nil")
+				return
+			}
+			require.NotNil(t, got, "resolved location should not be nil")
+			assert.True(t, tt.want.ConnectionIP.Equal(got.ConnectionIP), "connection IP should match")
+			assert.Equal(t, tt.want.CountryCode, got.CountryCode, "country code should match")
+			assert.Equal(t, tt.want.CityName, got.CityName, "city name should match")
+			assert.Equal(t, tt.want.GeoNameID, got.GeoNameID, "geoname id should match")
+		})
+	}
+}

management/server/posture/affects_posture_test.go

@@ -0,0 +1,202 @@
+package posture
+
+import (
+	"context"
+	"net"
+	"net/netip"
+	"testing"
+
+	"github.com/stretchr/testify/assert"
+
+	nbpeer "github.com/netbirdio/netbird/management/server/peer"
+)
+
+// diffFrom builds a MetaDiff from the old/new snapshots AffectsPosture replays against.
+func diffFrom(oldMeta, newMeta nbpeer.PeerSystemMeta, oldLoc, newLoc nbpeer.Location) *nbpeer.MetaDiff {
+	return &nbpeer.MetaDiff{
+		OldMeta:     oldMeta,
+		NewMeta:     newMeta,
+		OldLocation: oldLoc,
+		NewLocation: newLoc,
+	}
+}
+
+func checks(def ChecksDefinition) []*Checks {
+	return []*Checks{{Checks: def}}
+}
+
+func TestAffectsPosture_NilDiff(t *testing.T) {
+	assert.False(t, AffectsPosture(context.Background(), nil, checks(ChecksDefinition{
+		NBVersionCheck: &NBVersionCheck{MinVersion: "1.0.0"},
+	})))
+}
+
+func TestAffectsPosture_NBVersion(t *testing.T) {
+	c := checks(ChecksDefinition{NBVersionCheck: &NBVersionCheck{MinVersion: "1.2.0"}})
+
+	tests := []struct {
+		name           string
+		oldVer, newVer string
+		want           bool
+	}{
+		{"both above min, no flip", "1.3.0", "1.4.0", false},
+		{"both below min, no flip", "1.0.0", "1.1.0", false},
+		{"crosses up below->above", "1.1.0", "1.3.0", true},
+		{"crosses down above->below", "1.3.0", "1.1.0", true},
+		{"unparsable old only -> flip", "garbage", "1.3.0", true},
+		{"unparsable both -> no flip", "garbage", "junk", false},
+	}
+	for _, tt := range tests {
+		t.Run(tt.name, func(t *testing.T) {
+			diff := diffFrom(
+				nbpeer.PeerSystemMeta{WtVersion: tt.oldVer},
+				nbpeer.PeerSystemMeta{WtVersion: tt.newVer},
+				nbpeer.Location{}, nbpeer.Location{},
+			)
+			assert.Equal(t, tt.want, AffectsPosture(context.Background(), diff, c))
+		})
+	}
+}
+
+func TestAffectsPosture_OSVersion_KernelBumpWithinMin(t *testing.T) {
+	c := checks(ChecksDefinition{OSVersionCheck: &OSVersionCheck{
+		Linux: &MinKernelVersionCheck{MinKernelVersion: "5.0.0"},
+	}})
+
+	// Kernel moves but stays above the minimum: verdict stays pass -> not affected.
+	withinMin := diffFrom(
+		nbpeer.PeerSystemMeta{GoOS: "linux", KernelVersion: "5.10.0-arch1"},
+		nbpeer.PeerSystemMeta{GoOS: "linux", KernelVersion: "5.15.0-arch2"},
+		nbpeer.Location{}, nbpeer.Location{},
+	)
+	assert.False(t, AffectsPosture(context.Background(), withinMin, c))
+
+	// Kernel drops below the minimum: verdict flips pass -> fail -> affected.
+	crossesDown := diffFrom(
+		nbpeer.PeerSystemMeta{GoOS: "linux", KernelVersion: "5.10.0-arch1"},
+		nbpeer.PeerSystemMeta{GoOS: "linux", KernelVersion: "4.19.0-arch1"},
+		nbpeer.Location{}, nbpeer.Location{},
+	)
+	assert.True(t, AffectsPosture(context.Background(), crossesDown, c))
+}
+
+func TestAffectsPosture_OSVersion_GoOSSwitchFlipsVerdict(t *testing.T) {
+	// Only Linux is constrained. An OS outside the switch (freebsd) passes; switching to a
+	// failing linux kernel flips the verdict pass -> fail.
+	c := checks(ChecksDefinition{OSVersionCheck: &OSVersionCheck{
+		Linux: &MinKernelVersionCheck{MinKernelVersion: "6.0.0"},
+	}})
+
+	diff := diffFrom(
+		nbpeer.PeerSystemMeta{GoOS: "freebsd"},
+		nbpeer.PeerSystemMeta{GoOS: "linux", KernelVersion: "4.19.0"},
+		nbpeer.Location{}, nbpeer.Location{},
+	)
+	assert.True(t, AffectsPosture(context.Background(), diff, c))
+}
+
+func TestAffectsPosture_Process_GoOSSwitchFlipsVerdict(t *testing.T) {
+	// Process runs at a linux path. Switching GoOS to windows (no WindowsPath configured)
+	// flips the verdict.
+	c := checks(ChecksDefinition{ProcessCheck: &ProcessCheck{
+		Processes: []Process{{LinuxPath: "/usr/bin/foo"}},
+	}})
+
+	files := []nbpeer.File{{Path: "/usr/bin/foo", ProcessIsRunning: true}}
+	diff := diffFrom(
+		nbpeer.PeerSystemMeta{GoOS: "linux", Files: files},
+		nbpeer.PeerSystemMeta{GoOS: "windows", Files: files},
+		nbpeer.Location{}, nbpeer.Location{},
+	)
+	assert.True(t, AffectsPosture(context.Background(), diff, c))
+}
+
+func TestAffectsPosture_Process_UnrelatedFileChange(t *testing.T) {
+	// A tracked process stays running while an unrelated file is added: the verdict does
+	// not move, so posture is not affected.
+	c := checks(ChecksDefinition{ProcessCheck: &ProcessCheck{
+		Processes: []Process{{LinuxPath: "/usr/bin/foo"}},
+	}})
+
+	diff := diffFrom(
+		nbpeer.PeerSystemMeta{GoOS: "linux", Files: []nbpeer.File{
+			{Path: "/usr/bin/foo", ProcessIsRunning: true},
+		}},
+		nbpeer.PeerSystemMeta{GoOS: "linux", Files: []nbpeer.File{
+			{Path: "/usr/bin/foo", ProcessIsRunning: true},
+			{Path: "/usr/bin/bar", ProcessIsRunning: true},
+		}},
+		nbpeer.Location{}, nbpeer.Location{},
+	)
+	assert.False(t, AffectsPosture(context.Background(), diff, c))
+}
+
+func TestAffectsPosture_GeoLocation(t *testing.T) {
+	c := checks(ChecksDefinition{GeoLocationCheck: &GeoLocationCheck{
+		Action:    CheckActionAllow,
+		Locations: []Location{{CountryCode: "DE"}},
+	}})
+
+	// Moving within allowed countries keeps the verdict; moving out flips it.
+	stayAllowed := diffFrom(
+		nbpeer.PeerSystemMeta{}, nbpeer.PeerSystemMeta{},
+		nbpeer.Location{CountryCode: "DE", CityName: "Berlin"},
+		nbpeer.Location{CountryCode: "DE", CityName: "Munich"},
+	)
+	assert.False(t, AffectsPosture(context.Background(), stayAllowed, c))
+
+	moveOut := diffFrom(
+		nbpeer.PeerSystemMeta{}, nbpeer.PeerSystemMeta{},
+		nbpeer.Location{CountryCode: "DE"},
+		nbpeer.Location{CountryCode: "FR"},
+	)
+	assert.True(t, AffectsPosture(context.Background(), moveOut, c))
+}
+
+func TestAffectsPosture_PeerNetworkRange_ConnectionIP(t *testing.T) {
+	// The check reads the connection IP. Moving out of the allowed range flips the verdict;
+	// moving within it does not.
+	_, allowed, _ := net.ParseCIDR("10.0.0.0/8")
+	c := checks(ChecksDefinition{PeerNetworkRangeCheck: &PeerNetworkRangeCheck{
+		Action: CheckActionAllow,
+		Ranges: []netip.Prefix{netip.MustParsePrefix(allowed.String())},
+	}})
+
+	movesOutOfRange := diffFrom(
+		nbpeer.PeerSystemMeta{}, nbpeer.PeerSystemMeta{},
+		nbpeer.Location{ConnectionIP: net.ParseIP("10.1.2.3")},
+		nbpeer.Location{ConnectionIP: net.ParseIP("8.8.8.8")},
+	)
+	assert.True(t, AffectsPosture(context.Background(), movesOutOfRange, c))
+
+	staysInRange := diffFrom(
+		nbpeer.PeerSystemMeta{}, nbpeer.PeerSystemMeta{},
+		nbpeer.Location{ConnectionIP: net.ParseIP("10.1.2.3")},
+		nbpeer.Location{ConnectionIP: net.ParseIP("10.9.9.9")},
+	)
+	assert.False(t, AffectsPosture(context.Background(), staysInRange, c))
+}
+
+func TestAffectsPosture_IrrelevantFieldChange(t *testing.T) {
+	// Hostname changes but no check reads it: not affected even with checks present.
+	c := checks(ChecksDefinition{
+		NBVersionCheck:   &NBVersionCheck{MinVersion: "1.0.0"},
+		GeoLocationCheck: &GeoLocationCheck{Action: CheckActionAllow, Locations: []Location{{CountryCode: "DE"}}},
+	})
+
+	diff := diffFrom(
+		nbpeer.PeerSystemMeta{Hostname: "old", WtVersion: "1.5.0"},
+		nbpeer.PeerSystemMeta{Hostname: "new", WtVersion: "1.5.0"},
+		nbpeer.Location{CountryCode: "DE"}, nbpeer.Location{CountryCode: "DE"},
+	)
+	assert.False(t, AffectsPosture(context.Background(), diff, c))
+}
+
+func TestAffectsPosture_NoChecks(t *testing.T) {
+	diff := diffFrom(
+		nbpeer.PeerSystemMeta{WtVersion: "1.0.0"},
+		nbpeer.PeerSystemMeta{WtVersion: "2.0.0"},
+		nbpeer.Location{}, nbpeer.Location{},
+	)
+	assert.False(t, AffectsPosture(context.Background(), diff, nil))
+}

management/server/posture/checks.go

@@ -7,6 +7,7 @@ import (
 	"regexp"
 
 	"github.com/hashicorp/go-version"
+	log "github.com/sirupsen/logrus"
 
 	nbpeer "github.com/netbirdio/netbird/management/server/peer"
 	"github.com/netbirdio/netbird/shared/management/http/api"
@@ -52,34 +53,46 @@ type Checks struct {
 	Checks ChecksDefinition `gorm:"serializer:json"`
 }
 
-// AffectsPosture reports whether the peer metadata changes described by diff can
+// AffectsPosture reports whether the change in diff flips the verdict of any check. It
-// alter the outcome of any of the given posture checks. It maps each check kind to
+// replays each check against the peer's old and new state and compares verdicts, so a
-// the metadata fields it inspects, so an unrelated change (e.g. a hostname update)
+// change that moves a field but stays the right side of a threshold (e.g. a kernel bump
-// does not force a posture re-evaluation.
+// still above the minimum) does not force a re-evaluation. See verdictChanged for how an
-func AffectsPosture(diff *nbpeer.MetaDiff, checks []*Checks) bool {
+// evaluation error counts.
+func AffectsPosture(ctx context.Context, diff *nbpeer.MetaDiff, checks []*Checks) bool {
 	if diff == nil {
 		return false
 	}
+
+	oldPeer := nbpeer.Peer{Meta: diff.OldMeta, Location: diff.OldLocation}
+	newPeer := nbpeer.Peer{Meta: diff.NewMeta, Location: diff.NewLocation}
+
 	for _, c := range checks {
-		if c.Checks.ProcessCheck != nil && diff.Files {
+		for _, check := range c.GetChecks() {
-			return true
+			if verdictChanged(ctx, check, oldPeer, newPeer) {
-		}
+				return true
-		if c.Checks.OSVersionCheck != nil && (diff.OSVersion || diff.OS || diff.KernelVersion) {
+			}
-			return true
-		}
-		if c.Checks.NBVersionCheck != nil && diff.WtVersion {
-			return true
-		}
-		if c.Checks.GeoLocationCheck != nil && diff.LocationChanged {
-			return true
-		}
-		if c.Checks.PeerNetworkRangeCheck != nil && diff.NetworkAddresses {
-			return true
 		}
 	}
 	return false
 }
 
+// verdictChanged replays check against old and new state and reports whether the verdict
+// differs. Like callers, it treats an evaluation error as deny: two errors are the same
+// verdict (no change), an error on one side only is a flip.
+func verdictChanged(ctx context.Context, check Check, oldPeer, newPeer nbpeer.Peer) bool {
+	oldPass, oldErr := check.Check(ctx, oldPeer)
+	newPass, newErr := check.Check(ctx, newPeer)
+
+	oldVerdict := oldPass && (oldErr == nil)
+	newVerdict := newPass && (newErr == nil)
+	changed := oldVerdict != newVerdict
+
+	log.WithContext(ctx).Tracef("posture check %s replay: verdict %t -> %t (changed=%t), errs: %v -> %v",
+		check.Name(), oldVerdict, newVerdict, changed, oldErr, newErr)
+
+	return changed
+}
+
 // ChecksDefinition contains definition of actual check
 type ChecksDefinition struct {
 	NBVersionCheck        *NBVersionCheck        `json:",omitempty"`

management/server/posture_checks_test.go

@@ -489,6 +489,7 @@ func TestArePostureCheckChangesAffectPeers(t *testing.T) {
 
 	policy := &types.Policy{
 		AccountID: account.Id,
+		Enabled:   true,
 		Rules: []*types.PolicyRule{
 			{
 				Enabled:      true,

management/server/user.go

@@ -1059,8 +1059,8 @@ func (am *DefaultAccountManager) BuildUserInfosForAccount(ctx context.Context, a
 		if err != nil {
 			return nil, err
 		}
-		log.WithContext(ctx).Debugf("Got %d users from ExternalCache for account %s", len(usersFromIntegration), accountID)
+		log.WithContext(ctx).Tracef("Got %d users from ExternalCache for account %s", len(usersFromIntegration), accountID)
-		log.WithContext(ctx).Debugf("Got %d users from InternalCache for account %s", len(queriedUsers), accountID)
+		log.WithContext(ctx).Tracef("Got %d users from InternalCache for account %s", len(queriedUsers), accountID)
 		queriedUsers = append(queriedUsers, usersFromIntegration...)
 	}
 

shared/management/status/error.go

@@ -48,6 +48,10 @@ type Type int32
 var (
 	ErrExtraSettingsNotFound = errors.New("extra settings not found")
 	ErrPeerAlreadyLoggedIn   = errors.New("peer with the same public key is already logged in")
+
+	// ErrNoAuthMethodProvided is returned when a peer login attempt carries neither a
+	// setup key nor an SSO token. Match it with errors.Is.
+	ErrNoAuthMethodProvided = Errorf(Unauthenticated, "no peer auth method provided, please use a setup key or interactive SSO login")
 )
 
 // Error is an internal error
@@ -66,6 +70,16 @@ func (e *Error) Error() string {
 	return e.Message
 }
 
+// Is reports whether target is an *Error with the same type and message,
+// enabling matching with errors.Is against sentinel errors.
+func (e *Error) Is(target error) bool {
+	var t *Error
+	if !errors.As(target, &t) {
+		return false
+	}
+	return e.ErrorType == t.ErrorType && e.Message == t.Message
+}
+
 // Errorf returns Error(ErrorType, fmt.Sprintf(format, a...)).
 func Errorf(errorType Type, format string, a ...interface{}) error {
 	return &Error{

shared/signal/client/grpc.go

@@ -78,6 +78,13 @@ type GrpcClient struct {
 	// transport-alive but no longer delivering messages. It is the source of
 	// truth IsHealthy reads, and is cleared once any frame is received again.
 	receiveStalled atomic.Bool
+	// receiveHandoffBlocked is set while the receive loop is parked handing a
+	// message to a busy decryption worker. The loop stops calling Recv (and
+	// markReceived) in that window, so the stream looks silent though it is
+	// healthy. The watchdog reads this to avoid misreading self-inflicted
+	// receive backpressure as a dead stream: reconnecting cannot help, since the
+	// new stream feeds the same worker, and only triggers a reconnect storm.
+	receiveHandoffBlocked atomic.Bool
 }
 
 // NewClient creates a new Signal client
@@ -439,6 +446,16 @@ func (c *GrpcClient) idleSinceReceive() time.Duration {
 	return time.Since(time.Unix(0, c.lastReceived.Load()))
 }
 
+// receiveAlive reports whether the receive stream shows liveness: it delivered a
+// frame within the inactivity threshold, or the receive loop is currently parked
+// handing a message to a busy decryption worker. In the latter case the loop has
+// stopped calling Recv, so the stream looks silent while being healthy, and
+// reconnecting would not help, so the watchdog must treat it as alive.
+func (c *GrpcClient) receiveAlive() bool {
+	return c.idleSinceReceive() < receiveInactivityThreshold ||
+		c.receiveHandoffBlocked.Load()
+}
+
 // watchReceiveStream guards against a receive stream that is transport-alive but
 // no longer delivering messages. While the stream is idle past
 // receiveInactivityThreshold it sends a self-addressed probe that the Signal
@@ -455,7 +472,7 @@ func (c *GrpcClient) watchReceiveStream(ctx context.Context, cancelStream contex
 		case <-ctx.Done():
 			return
 		case <-ticker.C:
-			if c.idleSinceReceive() < receiveInactivityThreshold {
+			if c.receiveAlive() {
 				probeSentAt = time.Time{}
 				continue
 			}
@@ -517,9 +534,14 @@ func (c *GrpcClient) receive(stream proto.SignalExchange_ConnectStreamClient) er
 			continue
 		}
 
+		// The handoff blocks while the worker is busy, which parks this loop and
+		// stops Recv. Flag it so the watchdog does not read the resulting silence
+		// as a dead stream.
+		c.receiveHandoffBlocked.Store(true)
 		if err := c.decryptionWorker.AddMsg(c.ctx, msg); err != nil {
 			log.Errorf("failed to add message to decryption worker: %v", err)
 		}
+		c.receiveHandoffBlocked.Store(false)
 	}
 }
 

shared/signal/client/watchdog_test.go

@@ -82,3 +82,27 @@ func TestReceiveProbeRoundTrips(t *testing.T) {
 		t.Fatal("self-addressed heartbeat did not round-trip back through the signal server")
 	}
 }
+
+// TestReceiveAliveTreatsHandoffBlockAsLiveness reproduces the false positive
+// where a busy decryption worker parks the receive loop on the worker handoff,
+// so Recv (and markReceived) stops firing even though the stream is healthy.
+// With the receive stream silent past the inactivity threshold but the loop
+// blocked on handoff, the watchdog must consider the stream alive rather than
+// tear it down (reconnecting feeds the same worker and would not help).
+func TestReceiveAliveTreatsHandoffBlockAsLiveness(t *testing.T) {
+	c := &GrpcClient{}
+
+	// Receive stream silent and the loop not blocked on handoff: genuinely stalled.
+	c.lastReceived.Store(time.Now().Add(-2 * receiveInactivityThreshold).UnixNano())
+	require.False(t, c.receiveAlive(), "silent stream with the receive loop idle must be treated as stalled")
+
+	// Receive stream silent but the loop is parked handing a message to a busy
+	// worker: self-inflicted backpressure, not a dead stream, must not tear down.
+	c.receiveHandoffBlocked.Store(true)
+	require.True(t, c.receiveAlive(), "a receive loop blocked on worker handoff must keep the stream alive")
+
+	// Handoff drained, loop back to reading, a frame just arrived: alive via the receive path.
+	c.receiveHandoffBlocked.Store(false)
+	c.markReceived()
+	require.True(t, c.receiveAlive(), "a freshly received frame must keep the stream alive")
+}

sharedsock/sock_linux_test.go

@@ -1,3 +1,5 @@
+//go:build privileged
+
 package sharedsock
 
 import (