[client] add profiling dumps to debug package (#3517)

enhances debugging capabilities by adding support for goroutine, mutex, and block profiling while updating state dump tracking and refining test and release settings.

- Adds pprof-based profiling for goroutine, mutex, and block profiles in the debug bundle.
- Updates state dump functionality by incorporating new status and key fields.
- Adjusts test validations and default flag/retention settings.

.github/workflows/release.yml

@@ -87,25 +87,25 @@ jobs:
         with:
           name: release
           path: dist/
-          retention-days: 3
+          retention-days: 7
       - name: upload linux packages
         uses: actions/upload-artifact@v4
         with:
           name: linux-packages
           path: dist/netbird_linux**
-          retention-days: 3
+          retention-days: 7
       - name: upload windows packages
         uses: actions/upload-artifact@v4
         with:
           name: windows-packages
           path: dist/netbird_windows**
-          retention-days: 3
+          retention-days: 7
       - name: upload macos packages
         uses: actions/upload-artifact@v4
         with:
           name: macos-packages
           path: dist/netbird_darwin**
-          retention-days: 3
+          retention-days: 7
 
   release_ui:
     runs-on: ubuntu-latest

client/cmd/root.go

@@ -180,7 +180,7 @@ func init() {
 	upCmd.PersistentFlags().BoolVar(&serverSSHAllowed, serverSSHAllowedFlag, false, "Allow SSH server on peer. If enabled, the SSH server will be permitted")
 	upCmd.PersistentFlags().BoolVar(&autoConnectDisabled, disableAutoConnectFlag, false, "Disables auto-connect feature. If enabled, then the client won't connect automatically when the service starts.")
 
-	debugCmd.PersistentFlags().BoolVarP(&debugSystemInfoFlag, systemInfoFlag, "S", false, "Adds system information to the debug bundle")
+	debugCmd.PersistentFlags().BoolVarP(&debugSystemInfoFlag, systemInfoFlag, "S", true, "Adds system information to the debug bundle")
 }
 
 // SetupCloseHandler handles SIGTERM signal and exits with success

client/internal/peer/conn.go

@@ -140,7 +140,7 @@ func NewConn(engineCtx context.Context, config ConnConfig, statusRecorder *Statu
 		statusRelay:    NewAtomicConnStatus(),
 		statusICE:      NewAtomicConnStatus(),
 		semaphore:      semaphore,
-		dumpState:      newStateDump(connLog),
+		dumpState:      newStateDump(config.Key, connLog, statusRecorder),
 	}
 
 	ctrl := isController(config)
@@ -258,7 +258,7 @@ func (conn *Conn) Close() {
 // doesn't block, discards the message if connection wasn't ready
 func (conn *Conn) OnRemoteAnswer(answer OfferAnswer) bool {
 	conn.dumpState.RemoteAnswer()
-	conn.log.Infof("OnRemoteAnswer, status ICE: %s, status relay: %s", conn.statusICE, conn.statusRelay)
+	conn.log.Infof("OnRemoteAnswer, priority: %s, status ICE: %s, status relay: %s", conn.currentConnPriority, conn.statusICE, conn.statusRelay)
 	return conn.handshaker.OnRemoteAnswer(answer)
 }
 

client/internal/peer/state_dump.go

@@ -10,6 +10,8 @@ import (
 
 type stateDump struct {
 	log    *log.Entry
+	status *Status
+	key    string
 
 	sentOffer       int
 	remoteOffer     int
@@ -24,9 +26,11 @@ type stateDump struct {
 	mu sync.Mutex
 }
 
-func newStateDump(log *log.Entry) *stateDump {
+func newStateDump(key string, log *log.Entry, statusRecorder *Status) *stateDump {
 	return &stateDump{
 		log:    log,
+		status: statusRecorder,
+		key:    key,
 	}
 }
 
@@ -66,8 +70,14 @@ func (s *stateDump) dumpState() {
 	s.mu.Lock()
 	defer s.mu.Unlock()
 
-	s.log.Infof("Dump stat: SentOffer: %d, RemoteOffer: %d, RemoteAnswer: %d, RemoteCandidate: %d, P2PConnected: %d, SwitchToRelay: %d, WGCheckSuccess: %d, RelayConnected: %d, LocalProxies: %d",
-		s.sentOffer, s.remoteOffer, s.remoteAnswer, s.remoteCandidate, s.p2pConnected, s.switchToRelay, s.wgCheckSuccess, s.relayConnected, s.localProxies)
+	status := "unknown"
+	state, e := s.status.GetPeer(s.key)
+	if e == nil {
+		status = state.ConnStatus.String()
+	}
+
+	s.log.Infof("Dump stat: Status: %s, SentOffer: %d, RemoteOffer: %d, RemoteAnswer: %d, RemoteCandidate: %d, P2PConnected: %d, SwitchToRelay: %d, WGCheckSuccess: %d, RelayConnected: %d, LocalProxies: %d",
+		status, s.sentOffer, s.remoteOffer, s.remoteAnswer, s.remoteCandidate, s.p2pConnected, s.switchToRelay, s.wgCheckSuccess, s.relayConnected, s.localProxies)
 }
 
 func (s *stateDump) RemoteAnswer() {

client/internal/peer/wg_watcher_test.go

@@ -43,7 +43,7 @@ func TestWGWatcher_EnableWgWatcher(t *testing.T) {
 
 	mlog := log.WithField("peer", "tet")
 	mocWgIface := &MocWgIface{}
-	watcher := NewWGWatcher(mlog, mocWgIface, "", newStateDump(mlog))
+	watcher := NewWGWatcher(mlog, mocWgIface, "", newStateDump("peer", mlog, &Status{}))
 
 	ctx, cancel := context.WithCancel(context.Background())
 	defer cancel()
@@ -72,7 +72,7 @@ func TestWGWatcher_ReEnable(t *testing.T) {
 
 	mlog := log.WithField("peer", "tet")
 	mocWgIface := &MocWgIface{}
-	watcher := NewWGWatcher(mlog, mocWgIface, "", newStateDump(mlog))
+	watcher := NewWGWatcher(mlog, mocWgIface, "", newStateDump("peer", mlog, &Status{}))
 
 	ctx, cancel := context.WithCancel(context.Background())
 	defer cancel()

client/server/debug.go

@@ -17,6 +17,7 @@ import (
 	"os"
 	"path/filepath"
 	"runtime"
+	"runtime/pprof"
 	"sort"
 	"strings"
 	"time"
@@ -46,6 +47,9 @@ nftables.txt: Anonymized nftables rules with packet counters, if --system-info f
 config.txt: Anonymized configuration information of the NetBird client.
 network_map.json: Anonymized network map containing peer configurations, routes, DNS settings, and firewall rules.
 state.json: Anonymized client state dump containing netbird states.
+mutex.prof: Mutex profiling information.
+goroutine.prof: Goroutine profiling information.
+block.prof: Block profiling information.
 
 
 Anonymization Process
@@ -88,6 +92,14 @@ The state file follows the same anonymization rules as other files:
 - Domain names are consistently anonymized
 - Technical identifiers and non-sensitive data remain unchanged
 
+Mutex, Goroutines, and Block Profiling Files
+The goroutine, block, and mutex profiling files contains process information that might help the NetBird team diagnose performance issues. The information in these files don't contain personal data.
+You can check each using the following go command:
+
+go tool pprof -http=:8088 mutex.prof
+
+This will open a web browser tab with the profiling information.
+
 Routes
 For anonymized routes, the IP addresses are replaced as described above. The prefix length remains unchanged. Note that for prefixes, the anonymized IP might not be a network address, but the prefix length is still correct.
 
@@ -188,6 +200,10 @@ func (s *Server) createArchive(bundlePath *os.File, req *proto.DebugBundleReques
 		s.addSystemInfo(req, anonymizer, archive)
 	}
 
+	if err := s.addProf(req, anonymizer, archive); err != nil {
+		log.Errorf("Failed to add goroutines rules to debug bundle: %v", err)
+	}
+
 	if err := s.addNetworkMap(req, anonymizer, archive); err != nil {
 		return fmt.Errorf("add network map: %w", err)
 	}
@@ -310,6 +326,29 @@ func (s *Server) addCommonConfigFields(configContent *strings.Builder) {
 	configContent.WriteString(fmt.Sprintf("BlockLANAccess: %v\n", s.config.BlockLANAccess))
 }
 
+func (s *Server) addProf(req *proto.DebugBundleRequest, anonymizer *anonymize.Anonymizer, archive *zip.Writer) error {
+	runtime.SetBlockProfileRate(1)
+	_ = runtime.SetMutexProfileFraction(1)
+	defer runtime.SetBlockProfileRate(0)
+	defer runtime.SetMutexProfileFraction(0)
+
+	time.Sleep(5 * time.Second)
+
+	for _, profile := range []string{"goroutine", "block", "mutex"} {
+		var buff []byte
+		myBuff := bytes.NewBuffer(buff)
+		err := pprof.Lookup(profile).WriteTo(myBuff, 0)
+		if err != nil {
+			return fmt.Errorf("write %s profile: %w", profile, err)
+		}
+
+		if err := addFileToZip(archive, myBuff, profile+".prof"); err != nil {
+			return fmt.Errorf("add %s file to zip: %w", profile, err)
+		}
+	}
+	return nil
+}
+
 func (s *Server) addRoutes(req *proto.DebugBundleRequest, anonymizer *anonymize.Anonymizer, archive *zip.Writer) error {
 	routes, err := systemops.GetRoutesFromTable()
 	if err != nil {

management/server/store/sql_store.go

@@ -220,6 +220,10 @@ func generateAccountSQLTypes(account *types.Account) {
 		account.SetupKeysG = append(account.SetupKeysG, *key)
 	}
 
+	if len(account.SetupKeys) != len(account.SetupKeysG) {
+		log.Warnf("SetupKeysG length mismatch for account %s", account.Id)
+	}
+
 	for id, peer := range account.Peers {
 		peer.ID = id
 		account.PeersG = append(account.PeersG, *peer)

management/server/store/sql_store_test.go

@@ -148,6 +148,10 @@ func runLargeTest(t *testing.T, store Store) {
 		account.NameServerGroups[nameserver.ID] = nameserver
 
 		setupKey, _ := types.GenerateDefaultSetupKey()
+		_, exists := account.SetupKeys[setupKey.Key]
+		if exists {
+			t.Errorf("setup key already exists")
+		}
 		account.SetupKeys[setupKey.Key] = setupKey
 	}