Pass the new data down from the websocket

fix: allow empty config file bootstrap before provisioning

.github/CODEOWNERS

@@ -0,0 +1 @@
+* @oschwartz10612 @miloschwartz

blueprint.yaml

@@ -1,37 +0,0 @@
-resources:
-  resource-nice-id:
-    name: this is my resource
-    protocol: http
-    full-domain: level1.test3.example.com
-    host-header: example.com
-    tls-server-name: example.com
-    auth:
-      pincode: 123456
-      password: sadfasdfadsf
-      sso-enabled: true
-      sso-roles:
-      - Member
-      sso-users:
-      - owen@pangolin.net
-      whitelist-users:
-      - owen@pangolin.net
-    targets:
-    # - site: glossy-plains-viscacha-rat
-    - hostname: localhost
-      method: http
-      port: 8000
-      healthcheck:
-        port: 8000
-        hostname: localhost
-    # - site: glossy-plains-viscacha-rat
-    - hostname: localhost
-      method: http
-      port: 8001
-  resource-nice-id2:
-    name: this is other resource
-    protocol: tcp
-    proxy-port: 3000
-    targets:
-    # - site: glossy-plains-viscacha-rat
-    - hostname: localhost
-      port: 3000

clients/clients.go

@@ -40,12 +40,17 @@ type WgConfig struct {
 }
 
 type Target struct {
-	SourcePrefix   string      `json:"sourcePrefix"`
-	SourcePrefixes []string    `json:"sourcePrefixes"`
-	DestPrefix     string      `json:"destPrefix"`
-	RewriteTo      string      `json:"rewriteTo,omitempty"`
-	DisableIcmp    bool        `json:"disableIcmp,omitempty"`
-	PortRange      []PortRange `json:"portRange,omitempty"`
+	SourcePrefix   string                 `json:"sourcePrefix"`
+	SourcePrefixes []string               `json:"sourcePrefixes"`
+	DestPrefix     string                 `json:"destPrefix"`
+	RewriteTo      string                 `json:"rewriteTo,omitempty"`
+	DisableIcmp    bool                   `json:"disableIcmp,omitempty"`
+	PortRange      []PortRange            `json:"portRange,omitempty"`
+	ResourceId     int                    `json:"resourceId,omitempty"`
+	Protocol       string                 `json:"protocol,omitempty"`    // for now practicably either http or https
+	HTTPTargets    []netstack2.HTTPTarget `json:"httpTargets,omitempty"` // for http protocol, list of downstream services to load balance across
+	TLSCert        string                 `json:"tlsCert,omitempty"`     // PEM-encoded certificate for incoming HTTPS termination
+	TLSKey         string                 `json:"tlsKey,omitempty"`      // PEM-encoded private key for incoming HTTPS termination
 }
 
 type PortRange struct {
@@ -73,18 +78,18 @@ type PeerReading struct {
 }
 
 type WireGuardService struct {
-	interfaceName string
-	mtu           int
-	client        *websocket.Client
-	config        WgConfig
-	key           wgtypes.Key
-	newtId        string
-	lastReadings  map[string]PeerReading
-	mu            sync.Mutex
-	Port          uint16
-	host          string
-	serverPubKey  string
-	token         string
+	interfaceName        string
+	mtu                  int
+	client               *websocket.Client
+	config               WgConfig
+	key                  wgtypes.Key
+	newtId               string
+	lastReadings         map[string]PeerReading
+	mu                   sync.Mutex
+	Port                 uint16
+	host                 string
+	serverPubKey         string
+	token                string
 	stopGetConfig        func()
 	pendingConfigChainId string
 	// Netstack fields
@@ -206,6 +211,15 @@ func (s *WireGuardService) Close() {
 		s.stopGetConfig = nil
 	}
 
+	// Flush access logs before tearing down the tunnel
+	if s.tnet != nil {
+		if ph := s.tnet.GetProxyHandler(); ph != nil {
+			if al := ph.GetAccessLogger(); al != nil {
+				al.Close()
+			}
+		}
+	}
+
 	// Stop the direct UDP relay first
 	s.StopDirectUDPRelay()
 
@@ -687,7 +701,18 @@ func (s *WireGuardService) syncTargets(desiredTargets []Target) error {
 				})
 			}
 
-			s.tnet.AddProxySubnetRule(sourcePrefix, destPrefix, target.RewriteTo, portRanges, target.DisableIcmp)
+			s.tnet.AddProxySubnetRule(netstack2.SubnetRule{
+				SourcePrefix: sourcePrefix,
+				DestPrefix:   destPrefix,
+				RewriteTo:    target.RewriteTo,
+				PortRanges:   portRanges,
+				DisableIcmp:  target.DisableIcmp,
+				ResourceId:   target.ResourceId,
+				Protocol:     target.Protocol,
+				HTTPTargets:  target.HTTPTargets,
+				TLSCert:      target.TLSCert,
+				TLSKey:       target.TLSKey,
+			})
 			logger.Info("Added target %s -> %s during sync", target.SourcePrefix, target.DestPrefix)
 		}
 	}
@@ -818,6 +843,13 @@ func (s *WireGuardService) ensureWireguardInterface(wgconfig WgConfig) error {
 
 	s.TunnelIP = tunnelIP.String()
 
+	// Configure the access log sender to ship compressed session logs via websocket
+	s.tnet.SetAccessLogSender(func(data string) error {
+		return s.client.SendMessageNoLog("newt/access-log", map[string]interface{}{
+			"compressed": data,
+		})
+	})
+
 	// Create WireGuard device using the shared bind
 	s.device = device.NewDevice(s.tun, s.sharedBind, device.NewLogger(
 		device.LogLevelSilent, // Use silent logging by default - could be made configurable
@@ -938,7 +970,18 @@ func (s *WireGuardService) ensureTargets(targets []Target) error {
 			if err != nil {
 				return fmt.Errorf("invalid CIDR %s: %v", sp, err)
 			}
-			s.tnet.AddProxySubnetRule(sourcePrefix, destPrefix, target.RewriteTo, portRanges, target.DisableIcmp)
+			s.tnet.AddProxySubnetRule(netstack2.SubnetRule{
+				SourcePrefix: sourcePrefix,
+				DestPrefix:   destPrefix,
+				RewriteTo:    target.RewriteTo,
+				PortRanges:   portRanges,
+				DisableIcmp:  target.DisableIcmp,
+				ResourceId:   target.ResourceId,
+				Protocol:     target.Protocol,
+				HTTPTargets:  target.HTTPTargets,
+				TLSCert:      target.TLSCert,
+				TLSKey:       target.TLSKey,
+			})
 			logger.Info("Added target subnet from %s to %s rewrite to %s with port ranges: %v", sp, target.DestPrefix, target.RewriteTo, target.PortRange)
 		}
 	}
@@ -1331,7 +1374,18 @@ func (s *WireGuardService) handleAddTarget(msg websocket.WSMessage) {
 				logger.Info("Invalid CIDR %s: %v", sp, err)
 				continue
 			}
-			s.tnet.AddProxySubnetRule(sourcePrefix, destPrefix, target.RewriteTo, portRanges, target.DisableIcmp)
+			s.tnet.AddProxySubnetRule(netstack2.SubnetRule{
+				SourcePrefix: sourcePrefix,
+				DestPrefix:   destPrefix,
+				RewriteTo:    target.RewriteTo,
+				PortRanges:   portRanges,
+				DisableIcmp:  target.DisableIcmp,
+				ResourceId:   target.ResourceId,
+				Protocol:     target.Protocol,
+				HTTPTargets:  target.HTTPTargets,
+				TLSCert:      target.TLSCert,
+				TLSKey:       target.TLSKey,
+			})
 			logger.Info("Added target subnet from %s to %s rewrite to %s with port ranges: %v", sp, target.DestPrefix, target.RewriteTo, target.PortRange)
 		}
 	}
@@ -1449,7 +1503,18 @@ func (s *WireGuardService) handleUpdateTarget(msg websocket.WSMessage) {
 				logger.Info("Invalid CIDR %s: %v", sp, err)
 				continue
 			}
-			s.tnet.AddProxySubnetRule(sourcePrefix, destPrefix, target.RewriteTo, portRanges, target.DisableIcmp)
+			s.tnet.AddProxySubnetRule(netstack2.SubnetRule{
+				SourcePrefix: sourcePrefix,
+				DestPrefix:   destPrefix,
+				RewriteTo:    target.RewriteTo,
+				PortRanges:   portRanges,
+				DisableIcmp:  target.DisableIcmp,
+				ResourceId:   target.ResourceId,
+				Protocol:     target.Protocol,
+				HTTPTargets:  target.HTTPTargets,
+				TLSCert:      target.TLSCert,
+				TLSKey:       target.TLSKey,
+			})
 			logger.Info("Added target subnet from %s to %s rewrite to %s with port ranges: %v", sp, target.DestPrefix, target.RewriteTo, target.PortRange)
 		}
 	}

common.go

@@ -8,6 +8,7 @@ import (
 	"net"
 	"os"
 	"os/exec"
+	"regexp"
 	"strings"
 	"time"
 
@@ -516,15 +517,41 @@ func executeUpdownScript(action, proto, target string) (string, error) {
 	return target, nil
 }
 
-func sendBlueprint(client *websocket.Client) error {
-	if blueprintFile == "" {
+// interpolateBlueprint finds all {{...}} tokens in the raw blueprint bytes and
+// replaces recognised schemes with their resolved values. Currently supported:
+//
+//   - env.<VAR>  – replaced with the value of the named environment variable
+//
+// Any token that does not match a supported scheme is left as-is so that
+// future schemes (e.g. tag., api.) are preserved rather than silently dropped.
+func interpolateBlueprint(data []byte) []byte {
+	re := regexp.MustCompile(`\{\{([^}]+)\}\}`)
+	return re.ReplaceAllFunc(data, func(match []byte) []byte {
+		// strip the surrounding {{ }}
+		inner := strings.TrimSpace(string(match[2 : len(match)-2]))
+
+		if strings.HasPrefix(inner, "env.") {
+			varName := strings.TrimPrefix(inner, "env.")
+			return []byte(os.Getenv(varName))
+		}
+
+		// unrecognised scheme – leave the token untouched
+		return match
+	})
+}
+
+func sendBlueprint(client *websocket.Client, file string) error {
+	if file == "" {
 		return nil
 	}
 	// try to read the blueprint file
-	blueprintData, err := os.ReadFile(blueprintFile)
+	blueprintData, err := os.ReadFile(file)
 	if err != nil {
 		logger.Error("Failed to read blueprint file: %v", err)
 	} else {
+		// interpolate {{env.VAR}} (and any future schemes) before parsing
+		blueprintData = interpolateBlueprint(blueprintData)
+
 		// first we should convert the yaml to json and error if the yaml is bad
 		var yamlObj interface{}
 		var blueprintJsonData string

main.go

@@ -155,8 +155,9 @@ var (
 	region            string
 	metricsAsyncBytes bool
 	pprofEnabled      bool
-	blueprintFile     string
-	noCloud           bool
+	blueprintFile              string
+	provisioningBlueprintFile  string
+	noCloud                    bool
 
 	// New mTLS configuration variables
 	tlsClientCert string
@@ -165,6 +166,15 @@ var (
 
 	// Legacy PKCS12 support (deprecated)
 	tlsPrivateKey string
+
+	// Provisioning key – exchanged once for a permanent newt ID + secret
+	provisioningKey string
+
+	// Optional name for the site created during provisioning
+	newtName string
+
+	// Path to config file (overrides CONFIG_FILE env var and default location)
+	configFile string
 )
 
 // generateChainId generates a random chain ID for deduplicating round-trip messages.
@@ -275,8 +285,12 @@ func runNewtMain(ctx context.Context) {
 		tlsPrivateKey = os.Getenv("TLS_CLIENT_CERT")
 	}
 	blueprintFile = os.Getenv("BLUEPRINT_FILE")
+	provisioningBlueprintFile = os.Getenv("PROVISIONING_BLUEPRINT_FILE")
 	noCloudEnv := os.Getenv("NO_CLOUD")
 	noCloud = noCloudEnv == "true"
+	provisioningKey = os.Getenv("NEWT_PROVISIONING_KEY")
+	newtName = os.Getenv("NEWT_NAME")
+	configFile = os.Getenv("CONFIG_FILE")
 
 	if endpoint == "" {
 		flag.StringVar(&endpoint, "endpoint", "", "Endpoint of your pangolin server")
@@ -325,6 +339,15 @@ func runNewtMain(ctx context.Context) {
 	}
 	// load the prefer endpoint just as a flag
 	flag.StringVar(&preferEndpoint, "prefer-endpoint", "", "Prefer this endpoint for the connection (if set, will override the endpoint from the server)")
+	if provisioningKey == "" {
+		flag.StringVar(&provisioningKey, "provisioning-key", "", "One-time provisioning key used to obtain a newt ID and secret from the server")
+	}
+	if newtName == "" {
+		flag.StringVar(&newtName, "name", "", "Name for the site created during provisioning (supports {{env.VAR}} interpolation)")
+	}
+	if configFile == "" {
+		flag.StringVar(&configFile, "config-file", "", "Path to config file (overrides CONFIG_FILE env var and default location)")
+	}
 
 	// Add new mTLS flags
 	if tlsClientCert == "" {
@@ -372,6 +395,9 @@ func runNewtMain(ctx context.Context) {
 	if blueprintFile == "" {
 		flag.StringVar(&blueprintFile, "blueprint-file", "", "Path to blueprint file (if unset, no blueprint will be applied)")
 	}
+	if provisioningBlueprintFile == "" {
+		flag.StringVar(&provisioningBlueprintFile, "provisioning-blueprint-file", "", "Path to blueprint file applied once after a provisioning credential exchange (if unset, no provisioning blueprint will be applied)")
+	}
 	if noCloudEnv == "" {
 		flag.BoolVar(&noCloud, "no-cloud", false, "Disable cloud failover")
 	}
@@ -599,10 +625,20 @@ func runNewtMain(ctx context.Context) {
 		endpoint,
 		30*time.Second,
 		opt,
+		websocket.WithConfigFile(configFile),
 	)
 	if err != nil {
 		logger.Fatal("Failed to create client: %v", err)
 	}
+	// If a provisioning key was supplied via CLI / env and the config file did
+	// not already carry one, inject it now so provisionIfNeeded() can use it.
+	if provisioningKey != "" && client.GetConfig().ProvisioningKey == "" {
+		client.GetConfig().ProvisioningKey = provisioningKey
+	}
+	if newtName != "" && client.GetConfig().Name == "" {
+		client.GetConfig().Name = newtName
+	}
+
 	endpoint = client.GetConfig().Endpoint // Update endpoint from config
 	id = client.GetConfig().ID             // Update ID from config
 	// Update site labels for metrics with the resolved ID
@@ -1789,6 +1825,34 @@ persistent_keepalive_interval=5`, util.FixKey(privateKey.String()), util.FixKey(
 			} else {
 				logger.Warn("CLIENTS WILL NOT WORK ON THIS VERSION OF NEWT WITH THIS VERSION OF PANGOLIN, PLEASE UPDATE THE SERVER TO 1.13 OR HIGHER OR DOWNGRADE NEWT")
 			}
+			
+			sendBlueprint(client, blueprintFile)
+			if client.WasJustProvisioned() {
+				logger.Info("Provisioning detected – sending provisioning blueprint")
+				sendBlueprint(client, provisioningBlueprintFile)
+			}
+		} else {
+			// Resend current health check status for all targets in case the server
+			// missed updates while newt was disconnected.
+			targets := healthMonitor.GetTargets()
+			if len(targets) > 0 {
+				healthStatuses := make(map[int]interface{})
+				for id, target := range targets {
+					healthStatuses[id] = map[string]interface{}{
+						"status":     target.Status.String(),
+						"lastCheck":  target.LastCheck.Format(time.RFC3339),
+						"checkCount": target.CheckCount,
+						"lastError":  target.LastError,
+						"config":     target.Config,
+					}
+				}
+				logger.Debug("Reconnected: resending health check status for %d targets", len(healthStatuses))
+				if err := client.SendMessage("newt/healthcheck/status", map[string]interface{}{
+					"targets": healthStatuses,
+				}); err != nil {
+					logger.Error("Failed to resend health check status on reconnect: %v", err)
+				}
+			}
 		}
 
 		// Send registration message to the server for backward compatibility
@@ -1801,8 +1865,6 @@ persistent_keepalive_interval=5`, util.FixKey(privateKey.String()), util.FixKey(
 			"chainId":             bcChainId,
 		})
 
-		sendBlueprint(client)
-
 		if err != nil {
 			logger.Error("Failed to send registration message: %v", err)
 			return err

netstack2/access_log.go

@@ -0,0 +1,514 @@
+package netstack2
+
+import (
+	"bytes"
+	"compress/zlib"
+	"crypto/rand"
+	"encoding/base64"
+	"encoding/hex"
+	"encoding/json"
+	"net"
+	"sort"
+	"sync"
+	"time"
+
+	"github.com/fosrl/newt/logger"
+)
+
+const (
+	// flushInterval is how often the access logger flushes completed sessions to the server
+	flushInterval = 60 * time.Second
+
+	// maxBufferedSessions is the max number of completed sessions to buffer before forcing a flush
+	maxBufferedSessions = 100
+
+	// sessionGapThreshold is the maximum gap between the end of one connection
+	// and the start of the next for them to be considered part of the same session.
+	// If the gap exceeds this, a new consolidated session is created.
+	sessionGapThreshold = 5 * time.Second
+
+	// minConnectionsToConsolidate is the minimum number of connections in a group
+	// before we bother consolidating. Groups smaller than this are sent as-is.
+	minConnectionsToConsolidate = 2
+)
+
+// SendFunc is a callback that sends compressed access log data to the server.
+// The data is a base64-encoded zlib-compressed JSON array of AccessSession objects.
+type SendFunc func(data string) error
+
+// AccessSession represents a tracked access session through the proxy
+type AccessSession struct {
+	SessionID       string    `json:"sessionId"`
+	ResourceID      int       `json:"resourceId"`
+	SourceAddr      string    `json:"sourceAddr"`
+	DestAddr        string    `json:"destAddr"`
+	Protocol        string    `json:"protocol"`
+	StartedAt       time.Time `json:"startedAt"`
+	EndedAt         time.Time `json:"endedAt,omitempty"`
+	BytesTx         int64     `json:"bytesTx"`
+	BytesRx         int64     `json:"bytesRx"`
+	ConnectionCount int       `json:"connectionCount,omitempty"` // number of raw connections merged into this session (0 or 1 = single)
+}
+
+// udpSessionKey identifies a unique UDP "session" by src -> dst
+type udpSessionKey struct {
+	srcAddr  string
+	dstAddr  string
+	protocol string
+}
+
+// consolidationKey groups connections that may be part of the same logical session.
+// Source port is intentionally excluded so that many ephemeral-port connections
+// from the same source IP to the same destination are grouped together.
+type consolidationKey struct {
+	sourceIP   string // IP only, no port
+	destAddr   string // full host:port of the destination
+	protocol   string
+	resourceID int
+}
+
+// AccessLogger tracks access sessions for resources and periodically
+// flushes completed sessions to the server via a configurable SendFunc.
+type AccessLogger struct {
+	mu                sync.Mutex
+	sessions          map[string]*AccessSession       // active sessions: sessionID -> session
+	udpSessions       map[udpSessionKey]*AccessSession // active UDP sessions for dedup
+	completedSessions []*AccessSession                 // completed sessions waiting to be flushed
+	udpTimeout        time.Duration
+	sendFn            SendFunc
+	stopCh            chan struct{}
+	flushDone         chan struct{} // closed after the flush goroutine exits
+}
+
+// NewAccessLogger creates a new access logger.
+// udpTimeout controls how long a UDP session is kept alive without traffic before being ended.
+func NewAccessLogger(udpTimeout time.Duration) *AccessLogger {
+	al := &AccessLogger{
+		sessions:          make(map[string]*AccessSession),
+		udpSessions:       make(map[udpSessionKey]*AccessSession),
+		completedSessions: make([]*AccessSession, 0),
+		udpTimeout:        udpTimeout,
+		stopCh:            make(chan struct{}),
+		flushDone:         make(chan struct{}),
+	}
+	go al.backgroundLoop()
+	return al
+}
+
+// SetSendFunc sets the callback used to send compressed access log batches
+// to the server. This can be called after construction once the websocket
+// client is available.
+func (al *AccessLogger) SetSendFunc(fn SendFunc) {
+	al.mu.Lock()
+	defer al.mu.Unlock()
+	al.sendFn = fn
+}
+
+// generateSessionID creates a random session identifier
+func generateSessionID() string {
+	b := make([]byte, 8)
+	rand.Read(b)
+	return hex.EncodeToString(b)
+}
+
+// StartTCPSession logs the start of a TCP session and returns a session ID.
+func (al *AccessLogger) StartTCPSession(resourceID int, srcAddr, dstAddr string) string {
+	sessionID := generateSessionID()
+	now := time.Now()
+
+	session := &AccessSession{
+		SessionID:  sessionID,
+		ResourceID: resourceID,
+		SourceAddr: srcAddr,
+		DestAddr:   dstAddr,
+		Protocol:   "tcp",
+		StartedAt:  now,
+	}
+
+	al.mu.Lock()
+	al.sessions[sessionID] = session
+	al.mu.Unlock()
+
+	logger.Info("ACCESS START session=%s resource=%d proto=tcp src=%s dst=%s time=%s",
+		sessionID, resourceID, srcAddr, dstAddr, now.Format(time.RFC3339))
+
+	return sessionID
+}
+
+// EndTCPSession logs the end of a TCP session and queues it for sending.
+func (al *AccessLogger) EndTCPSession(sessionID string) {
+	now := time.Now()
+
+	al.mu.Lock()
+	session, ok := al.sessions[sessionID]
+	if ok {
+		session.EndedAt = now
+		delete(al.sessions, sessionID)
+		al.completedSessions = append(al.completedSessions, session)
+	}
+	shouldFlush := len(al.completedSessions) >= maxBufferedSessions
+	al.mu.Unlock()
+
+	if ok {
+		duration := now.Sub(session.StartedAt)
+		logger.Info("ACCESS END session=%s resource=%d proto=tcp src=%s dst=%s started=%s ended=%s duration=%s",
+			sessionID, session.ResourceID, session.SourceAddr, session.DestAddr,
+			session.StartedAt.Format(time.RFC3339), now.Format(time.RFC3339), duration)
+	}
+
+	if shouldFlush {
+		al.flush()
+	}
+}
+
+// TrackUDPSession starts or returns an existing UDP session. Returns the session ID.
+func (al *AccessLogger) TrackUDPSession(resourceID int, srcAddr, dstAddr string) string {
+	key := udpSessionKey{
+		srcAddr:  srcAddr,
+		dstAddr:  dstAddr,
+		protocol: "udp",
+	}
+
+	al.mu.Lock()
+	defer al.mu.Unlock()
+
+	if existing, ok := al.udpSessions[key]; ok {
+		return existing.SessionID
+	}
+
+	sessionID := generateSessionID()
+	now := time.Now()
+
+	session := &AccessSession{
+		SessionID:  sessionID,
+		ResourceID: resourceID,
+		SourceAddr: srcAddr,
+		DestAddr:   dstAddr,
+		Protocol:   "udp",
+		StartedAt:  now,
+	}
+
+	al.sessions[sessionID] = session
+	al.udpSessions[key] = session
+
+	logger.Info("ACCESS START session=%s resource=%d proto=udp src=%s dst=%s time=%s",
+		sessionID, resourceID, srcAddr, dstAddr, now.Format(time.RFC3339))
+
+	return sessionID
+}
+
+// EndUDPSession ends a UDP session and queues it for sending.
+func (al *AccessLogger) EndUDPSession(sessionID string) {
+	now := time.Now()
+
+	al.mu.Lock()
+	session, ok := al.sessions[sessionID]
+	if ok {
+		session.EndedAt = now
+		delete(al.sessions, sessionID)
+		key := udpSessionKey{
+			srcAddr:  session.SourceAddr,
+			dstAddr:  session.DestAddr,
+			protocol: "udp",
+		}
+		delete(al.udpSessions, key)
+		al.completedSessions = append(al.completedSessions, session)
+	}
+	shouldFlush := len(al.completedSessions) >= maxBufferedSessions
+	al.mu.Unlock()
+
+	if ok {
+		duration := now.Sub(session.StartedAt)
+		logger.Info("ACCESS END session=%s resource=%d proto=udp src=%s dst=%s started=%s ended=%s duration=%s",
+			sessionID, session.ResourceID, session.SourceAddr, session.DestAddr,
+			session.StartedAt.Format(time.RFC3339), now.Format(time.RFC3339), duration)
+	}
+
+	if shouldFlush {
+		al.flush()
+	}
+}
+
+// backgroundLoop handles periodic flushing and stale session reaping.
+func (al *AccessLogger) backgroundLoop() {
+	defer close(al.flushDone)
+
+	flushTicker := time.NewTicker(flushInterval)
+	defer flushTicker.Stop()
+
+	reapTicker := time.NewTicker(30 * time.Second)
+	defer reapTicker.Stop()
+
+	for {
+		select {
+		case <-al.stopCh:
+			return
+		case <-flushTicker.C:
+			al.flush()
+		case <-reapTicker.C:
+			al.reapStaleSessions()
+		}
+	}
+}
+
+// reapStaleSessions cleans up UDP sessions that were not properly ended.
+func (al *AccessLogger) reapStaleSessions() {
+	al.mu.Lock()
+	defer al.mu.Unlock()
+
+	staleThreshold := time.Now().Add(-5 * time.Minute)
+
+	for key, session := range al.udpSessions {
+		if session.StartedAt.Before(staleThreshold) && session.EndedAt.IsZero() {
+			now := time.Now()
+			session.EndedAt = now
+			duration := now.Sub(session.StartedAt)
+			logger.Info("ACCESS END (reaped) session=%s resource=%d proto=udp src=%s dst=%s started=%s ended=%s duration=%s",
+				session.SessionID, session.ResourceID, session.SourceAddr, session.DestAddr,
+				session.StartedAt.Format(time.RFC3339), now.Format(time.RFC3339), duration)
+			al.completedSessions = append(al.completedSessions, session)
+			delete(al.sessions, session.SessionID)
+			delete(al.udpSessions, key)
+		}
+	}
+}
+
+// extractIP strips the port from an address string and returns just the IP.
+// If the address has no port component it is returned as-is.
+func extractIP(addr string) string {
+	host, _, err := net.SplitHostPort(addr)
+	if err != nil {
+		// Might already be a bare IP
+		return addr
+	}
+	return host
+}
+
+// consolidateSessions takes a slice of completed sessions and merges bursts of
+// short-lived connections from the same source IP to the same destination into
+// single higher-level session entries.
+//
+// The algorithm:
+//  1. Group sessions by (sourceIP, destAddr, protocol, resourceID).
+//  2. Within each group, sort by StartedAt.
+//  3. Walk through the sorted list and merge consecutive sessions whose gap
+//     (previous EndedAt → next StartedAt) is ≤ sessionGapThreshold.
+//  4. For merged sessions the earliest StartedAt and latest EndedAt are kept,
+//     bytes are summed, and ConnectionCount records how many raw connections
+//     were folded in. If the merged connections used more than one source port,
+//     SourceAddr is set to just the IP (port omitted).
+//  5. Groups with fewer than minConnectionsToConsolidate members are passed
+//     through unmodified.
+func consolidateSessions(sessions []*AccessSession) []*AccessSession {
+	if len(sessions) <= 1 {
+		return sessions
+	}
+
+	// Group sessions by consolidation key
+	groups := make(map[consolidationKey][]*AccessSession)
+	for _, s := range sessions {
+		key := consolidationKey{
+			sourceIP:   extractIP(s.SourceAddr),
+			destAddr:   s.DestAddr,
+			protocol:   s.Protocol,
+			resourceID: s.ResourceID,
+		}
+		groups[key] = append(groups[key], s)
+	}
+
+	result := make([]*AccessSession, 0, len(sessions))
+
+	for key, group := range groups {
+		// Small groups don't need consolidation
+		if len(group) < minConnectionsToConsolidate {
+			result = append(result, group...)
+			continue
+		}
+
+		// Sort the group by start time so we can detect gaps
+		sort.Slice(group, func(i, j int) bool {
+			return group[i].StartedAt.Before(group[j].StartedAt)
+		})
+
+		// Walk through and merge runs that are within the gap threshold
+		var merged []*AccessSession
+		cur := cloneSession(group[0])
+		cur.ConnectionCount = 1
+		sourcePorts := make(map[string]struct{})
+		sourcePorts[cur.SourceAddr] = struct{}{}
+
+		for i := 1; i < len(group); i++ {
+			s := group[i]
+
+			// Determine the gap: from the latest end time we've seen so far to the
+			// start of the next connection.
+			gapRef := cur.EndedAt
+			if gapRef.IsZero() {
+				gapRef = cur.StartedAt
+			}
+			gap := s.StartedAt.Sub(gapRef)
+
+			if gap <= sessionGapThreshold {
+				// Merge into the current consolidated session
+				cur.ConnectionCount++
+				cur.BytesTx += s.BytesTx
+				cur.BytesRx += s.BytesRx
+				sourcePorts[s.SourceAddr] = struct{}{}
+
+				// Extend EndedAt to the latest time
+				endTime := s.EndedAt
+				if endTime.IsZero() {
+					endTime = s.StartedAt
+				}
+				if endTime.After(cur.EndedAt) {
+					cur.EndedAt = endTime
+				}
+			} else {
+				// Gap exceeded — finalize the current session and start a new one
+				finalizeMergedSourceAddr(cur, key.sourceIP, sourcePorts)
+				merged = append(merged, cur)
+
+				cur = cloneSession(s)
+				cur.ConnectionCount = 1
+				sourcePorts = make(map[string]struct{})
+				sourcePorts[s.SourceAddr] = struct{}{}
+			}
+		}
+
+		// Finalize the last accumulated session
+		finalizeMergedSourceAddr(cur, key.sourceIP, sourcePorts)
+		merged = append(merged, cur)
+
+		result = append(result, merged...)
+	}
+
+	return result
+}
+
+// cloneSession creates a shallow copy of an AccessSession.
+func cloneSession(s *AccessSession) *AccessSession {
+	cp := *s
+	return &cp
+}
+
+// finalizeMergedSourceAddr sets the SourceAddr on a consolidated session.
+// If multiple distinct source addresses (ports) were seen, the port is
+// stripped and only the IP is kept so the log isn't misleading.
+func finalizeMergedSourceAddr(s *AccessSession, sourceIP string, ports map[string]struct{}) {
+	if len(ports) > 1 {
+		// Multiple source ports — just report the IP
+		s.SourceAddr = sourceIP
+	}
+	// Otherwise keep the original SourceAddr which already has ip:port
+}
+
+// flush drains the completed sessions buffer, consolidates bursts of
+// short-lived connections, compresses with zlib, and sends via the SendFunc.
+func (al *AccessLogger) flush() {
+	al.mu.Lock()
+	if len(al.completedSessions) == 0 {
+		al.mu.Unlock()
+		return
+	}
+	batch := al.completedSessions
+	al.completedSessions = make([]*AccessSession, 0)
+	sendFn := al.sendFn
+	al.mu.Unlock()
+
+	if sendFn == nil {
+		logger.Debug("Access logger: no send function configured, discarding %d sessions", len(batch))
+		return
+	}
+
+	// Consolidate bursts of short-lived connections into higher-level sessions
+	originalCount := len(batch)
+	batch = consolidateSessions(batch)
+	if len(batch) != originalCount {
+		logger.Info("Access logger: consolidated %d raw connections into %d sessions", originalCount, len(batch))
+	}
+
+	compressed, err := compressSessions(batch)
+	if err != nil {
+		logger.Error("Access logger: failed to compress %d sessions: %v", len(batch), err)
+		return
+	}
+
+	if err := sendFn(compressed); err != nil {
+		logger.Error("Access logger: failed to send %d sessions: %v", len(batch), err)
+		// Re-queue the batch so we don't lose data
+		al.mu.Lock()
+		al.completedSessions = append(batch, al.completedSessions...)
+		// Cap re-queued data to prevent unbounded growth if server is unreachable
+		if len(al.completedSessions) > maxBufferedSessions*5 {
+			dropped := len(al.completedSessions) - maxBufferedSessions*5
+			al.completedSessions = al.completedSessions[:maxBufferedSessions*5]
+			logger.Warn("Access logger: buffer overflow, dropped %d oldest sessions", dropped)
+		}
+		al.mu.Unlock()
+		return
+	}
+
+	logger.Info("Access logger: sent %d sessions to server", len(batch))
+}
+
+// compressSessions JSON-encodes the sessions, compresses with zlib, and returns
+// a base64-encoded string suitable for embedding in a JSON message.
+func compressSessions(sessions []*AccessSession) (string, error) {
+	jsonData, err := json.Marshal(sessions)
+	if err != nil {
+		return "", err
+	}
+
+	var buf bytes.Buffer
+	w, err := zlib.NewWriterLevel(&buf, zlib.BestCompression)
+	if err != nil {
+		return "", err
+	}
+	if _, err := w.Write(jsonData); err != nil {
+		w.Close()
+		return "", err
+	}
+	if err := w.Close(); err != nil {
+		return "", err
+	}
+
+	return base64.StdEncoding.EncodeToString(buf.Bytes()), nil
+}
+
+// Close shuts down the background loop, ends all active sessions,
+// and performs one final flush to send everything to the server.
+func (al *AccessLogger) Close() {
+	// Signal the background loop to stop
+	select {
+	case <-al.stopCh:
+		// Already closed
+		return
+	default:
+		close(al.stopCh)
+	}
+
+	// Wait for the background loop to exit so we don't race on flush
+	<-al.flushDone
+
+	al.mu.Lock()
+	now := time.Now()
+
+	// End all active sessions and move them to the completed buffer
+	for _, session := range al.sessions {
+		if session.EndedAt.IsZero() {
+			session.EndedAt = now
+			duration := now.Sub(session.StartedAt)
+			logger.Info("ACCESS END (shutdown) session=%s resource=%d proto=%s src=%s dst=%s started=%s ended=%s duration=%s",
+				session.SessionID, session.ResourceID, session.Protocol, session.SourceAddr, session.DestAddr,
+				session.StartedAt.Format(time.RFC3339), now.Format(time.RFC3339), duration)
+			al.completedSessions = append(al.completedSessions, session)
+		}
+	}
+
+	al.sessions = make(map[string]*AccessSession)
+	al.udpSessions = make(map[udpSessionKey]*AccessSession)
+	al.mu.Unlock()
+
+	// Final flush to send all remaining sessions to the server
+	al.flush()
+}

netstack2/access_log_test.go

@@ -0,0 +1,811 @@
+package netstack2
+
+import (
+	"testing"
+	"time"
+)
+
+func TestExtractIP(t *testing.T) {
+	tests := []struct {
+		name     string
+		addr     string
+		expected string
+	}{
+		{"ipv4 with port", "192.168.1.1:12345", "192.168.1.1"},
+		{"ipv4 without port", "192.168.1.1", "192.168.1.1"},
+		{"ipv6 with port", "[::1]:12345", "::1"},
+		{"ipv6 without port", "::1", "::1"},
+		{"empty string", "", ""},
+	}
+
+	for _, tt := range tests {
+		t.Run(tt.name, func(t *testing.T) {
+			result := extractIP(tt.addr)
+			if result != tt.expected {
+				t.Errorf("extractIP(%q) = %q, want %q", tt.addr, result, tt.expected)
+			}
+		})
+	}
+}
+
+func TestConsolidateSessions_Empty(t *testing.T) {
+	result := consolidateSessions(nil)
+	if result != nil {
+		t.Errorf("expected nil, got %v", result)
+	}
+
+	result = consolidateSessions([]*AccessSession{})
+	if len(result) != 0 {
+		t.Errorf("expected empty slice, got %d items", len(result))
+	}
+}
+
+func TestConsolidateSessions_SingleSession(t *testing.T) {
+	now := time.Now()
+	sessions := []*AccessSession{
+		{
+			SessionID:  "abc123",
+			ResourceID: 1,
+			SourceAddr: "10.0.0.1:5000",
+			DestAddr:   "192.168.1.100:443",
+			Protocol:   "tcp",
+			StartedAt:  now,
+			EndedAt:    now.Add(1 * time.Second),
+		},
+	}
+
+	result := consolidateSessions(sessions)
+	if len(result) != 1 {
+		t.Fatalf("expected 1 session, got %d", len(result))
+	}
+	if result[0].SourceAddr != "10.0.0.1:5000" {
+		t.Errorf("expected source addr preserved, got %q", result[0].SourceAddr)
+	}
+}
+
+func TestConsolidateSessions_MergesBurstFromSameSourceIP(t *testing.T) {
+	now := time.Now()
+	sessions := []*AccessSession{
+		{
+			SessionID:  "s1",
+			ResourceID: 1,
+			SourceAddr: "10.0.0.1:5000",
+			DestAddr:   "192.168.1.100:443",
+			Protocol:   "tcp",
+			StartedAt:  now,
+			EndedAt:    now.Add(100 * time.Millisecond),
+			BytesTx:    100,
+			BytesRx:    200,
+		},
+		{
+			SessionID:  "s2",
+			ResourceID: 1,
+			SourceAddr: "10.0.0.1:5001",
+			DestAddr:   "192.168.1.100:443",
+			Protocol:   "tcp",
+			StartedAt:  now.Add(200 * time.Millisecond),
+			EndedAt:    now.Add(300 * time.Millisecond),
+			BytesTx:    150,
+			BytesRx:    250,
+		},
+		{
+			SessionID:  "s3",
+			ResourceID: 1,
+			SourceAddr: "10.0.0.1:5002",
+			DestAddr:   "192.168.1.100:443",
+			Protocol:   "tcp",
+			StartedAt:  now.Add(400 * time.Millisecond),
+			EndedAt:    now.Add(500 * time.Millisecond),
+			BytesTx:    50,
+			BytesRx:    75,
+		},
+	}
+
+	result := consolidateSessions(sessions)
+	if len(result) != 1 {
+		t.Fatalf("expected 1 consolidated session, got %d", len(result))
+	}
+
+	s := result[0]
+	if s.ConnectionCount != 3 {
+		t.Errorf("expected ConnectionCount=3, got %d", s.ConnectionCount)
+	}
+	if s.SourceAddr != "10.0.0.1" {
+		t.Errorf("expected source addr to be IP only (multiple ports), got %q", s.SourceAddr)
+	}
+	if s.DestAddr != "192.168.1.100:443" {
+		t.Errorf("expected dest addr preserved, got %q", s.DestAddr)
+	}
+	if s.StartedAt != now {
+		t.Errorf("expected StartedAt to be earliest time")
+	}
+	if s.EndedAt != now.Add(500*time.Millisecond) {
+		t.Errorf("expected EndedAt to be latest time")
+	}
+	expectedTx := int64(300)
+	expectedRx := int64(525)
+	if s.BytesTx != expectedTx {
+		t.Errorf("expected BytesTx=%d, got %d", expectedTx, s.BytesTx)
+	}
+	if s.BytesRx != expectedRx {
+		t.Errorf("expected BytesRx=%d, got %d", expectedRx, s.BytesRx)
+	}
+}
+
+func TestConsolidateSessions_SameSourcePortPreserved(t *testing.T) {
+	now := time.Now()
+	sessions := []*AccessSession{
+		{
+			SessionID:  "s1",
+			ResourceID: 1,
+			SourceAddr: "10.0.0.1:5000",
+			DestAddr:   "192.168.1.100:443",
+			Protocol:   "tcp",
+			StartedAt:  now,
+			EndedAt:    now.Add(100 * time.Millisecond),
+		},
+		{
+			SessionID:  "s2",
+			ResourceID: 1,
+			SourceAddr: "10.0.0.1:5000",
+			DestAddr:   "192.168.1.100:443",
+			Protocol:   "tcp",
+			StartedAt:  now.Add(200 * time.Millisecond),
+			EndedAt:    now.Add(300 * time.Millisecond),
+		},
+	}
+
+	result := consolidateSessions(sessions)
+	if len(result) != 1 {
+		t.Fatalf("expected 1 session, got %d", len(result))
+	}
+	if result[0].SourceAddr != "10.0.0.1:5000" {
+		t.Errorf("expected source addr with port preserved when all ports are the same, got %q", result[0].SourceAddr)
+	}
+	if result[0].ConnectionCount != 2 {
+		t.Errorf("expected ConnectionCount=2, got %d", result[0].ConnectionCount)
+	}
+}
+
+func TestConsolidateSessions_GapSplitsSessions(t *testing.T) {
+	now := time.Now()
+
+	// First burst
+	sessions := []*AccessSession{
+		{
+			SessionID:  "s1",
+			ResourceID: 1,
+			SourceAddr: "10.0.0.1:5000",
+			DestAddr:   "192.168.1.100:443",
+			Protocol:   "tcp",
+			StartedAt:  now,
+			EndedAt:    now.Add(100 * time.Millisecond),
+		},
+		{
+			SessionID:  "s2",
+			ResourceID: 1,
+			SourceAddr: "10.0.0.1:5001",
+			DestAddr:   "192.168.1.100:443",
+			Protocol:   "tcp",
+			StartedAt:  now.Add(200 * time.Millisecond),
+			EndedAt:    now.Add(300 * time.Millisecond),
+		},
+		// Big gap here (10 seconds)
+		{
+			SessionID:  "s3",
+			ResourceID: 1,
+			SourceAddr: "10.0.0.1:5002",
+			DestAddr:   "192.168.1.100:443",
+			Protocol:   "tcp",
+			StartedAt:  now.Add(10 * time.Second),
+			EndedAt:    now.Add(10*time.Second + 100*time.Millisecond),
+		},
+		{
+			SessionID:  "s4",
+			ResourceID: 1,
+			SourceAddr: "10.0.0.1:5003",
+			DestAddr:   "192.168.1.100:443",
+			Protocol:   "tcp",
+			StartedAt:  now.Add(10*time.Second + 200*time.Millisecond),
+			EndedAt:    now.Add(10*time.Second + 300*time.Millisecond),
+		},
+	}
+
+	result := consolidateSessions(sessions)
+	if len(result) != 2 {
+		t.Fatalf("expected 2 consolidated sessions (gap split), got %d", len(result))
+	}
+
+	// Find the sessions by their start time
+	var first, second *AccessSession
+	for _, s := range result {
+		if s.StartedAt.Equal(now) {
+			first = s
+		} else {
+			second = s
+		}
+	}
+
+	if first == nil || second == nil {
+		t.Fatal("could not find both consolidated sessions")
+	}
+
+	if first.ConnectionCount != 2 {
+		t.Errorf("first burst: expected ConnectionCount=2, got %d", first.ConnectionCount)
+	}
+	if second.ConnectionCount != 2 {
+		t.Errorf("second burst: expected ConnectionCount=2, got %d", second.ConnectionCount)
+	}
+}
+
+func TestConsolidateSessions_DifferentDestinationsNotMerged(t *testing.T) {
+	now := time.Now()
+	sessions := []*AccessSession{
+		{
+			SessionID:  "s1",
+			ResourceID: 1,
+			SourceAddr: "10.0.0.1:5000",
+			DestAddr:   "192.168.1.100:443",
+			Protocol:   "tcp",
+			StartedAt:  now,
+			EndedAt:    now.Add(100 * time.Millisecond),
+		},
+		{
+			SessionID:  "s2",
+			ResourceID: 1,
+			SourceAddr: "10.0.0.1:5001",
+			DestAddr:   "192.168.1.100:8080",
+			Protocol:   "tcp",
+			StartedAt:  now.Add(200 * time.Millisecond),
+			EndedAt:    now.Add(300 * time.Millisecond),
+		},
+	}
+
+	result := consolidateSessions(sessions)
+	// Each goes to a different dest port so they should not be merged
+	if len(result) != 2 {
+		t.Fatalf("expected 2 sessions (different destinations), got %d", len(result))
+	}
+}
+
+func TestConsolidateSessions_DifferentProtocolsNotMerged(t *testing.T) {
+	now := time.Now()
+	sessions := []*AccessSession{
+		{
+			SessionID:  "s1",
+			ResourceID: 1,
+			SourceAddr: "10.0.0.1:5000",
+			DestAddr:   "192.168.1.100:443",
+			Protocol:   "tcp",
+			StartedAt:  now,
+			EndedAt:    now.Add(100 * time.Millisecond),
+		},
+		{
+			SessionID:  "s2",
+			ResourceID: 1,
+			SourceAddr: "10.0.0.1:5001",
+			DestAddr:   "192.168.1.100:443",
+			Protocol:   "udp",
+			StartedAt:  now.Add(200 * time.Millisecond),
+			EndedAt:    now.Add(300 * time.Millisecond),
+		},
+	}
+
+	result := consolidateSessions(sessions)
+	if len(result) != 2 {
+		t.Fatalf("expected 2 sessions (different protocols), got %d", len(result))
+	}
+}
+
+func TestConsolidateSessions_DifferentResourceIDsNotMerged(t *testing.T) {
+	now := time.Now()
+	sessions := []*AccessSession{
+		{
+			SessionID:  "s1",
+			ResourceID: 1,
+			SourceAddr: "10.0.0.1:5000",
+			DestAddr:   "192.168.1.100:443",
+			Protocol:   "tcp",
+			StartedAt:  now,
+			EndedAt:    now.Add(100 * time.Millisecond),
+		},
+		{
+			SessionID:  "s2",
+			ResourceID: 2,
+			SourceAddr: "10.0.0.1:5001",
+			DestAddr:   "192.168.1.100:443",
+			Protocol:   "tcp",
+			StartedAt:  now.Add(200 * time.Millisecond),
+			EndedAt:    now.Add(300 * time.Millisecond),
+		},
+	}
+
+	result := consolidateSessions(sessions)
+	if len(result) != 2 {
+		t.Fatalf("expected 2 sessions (different resource IDs), got %d", len(result))
+	}
+}
+
+func TestConsolidateSessions_DifferentSourceIPsNotMerged(t *testing.T) {
+	now := time.Now()
+	sessions := []*AccessSession{
+		{
+			SessionID:  "s1",
+			ResourceID: 1,
+			SourceAddr: "10.0.0.1:5000",
+			DestAddr:   "192.168.1.100:443",
+			Protocol:   "tcp",
+			StartedAt:  now,
+			EndedAt:    now.Add(100 * time.Millisecond),
+		},
+		{
+			SessionID:  "s2",
+			ResourceID: 1,
+			SourceAddr: "10.0.0.2:5001",
+			DestAddr:   "192.168.1.100:443",
+			Protocol:   "tcp",
+			StartedAt:  now.Add(200 * time.Millisecond),
+			EndedAt:    now.Add(300 * time.Millisecond),
+		},
+	}
+
+	result := consolidateSessions(sessions)
+	if len(result) != 2 {
+		t.Fatalf("expected 2 sessions (different source IPs), got %d", len(result))
+	}
+}
+
+func TestConsolidateSessions_OutOfOrderInput(t *testing.T) {
+	now := time.Now()
+	// Provide sessions out of chronological order to verify sorting
+	sessions := []*AccessSession{
+		{
+			SessionID:  "s3",
+			ResourceID: 1,
+			SourceAddr: "10.0.0.1:5002",
+			DestAddr:   "192.168.1.100:443",
+			Protocol:   "tcp",
+			StartedAt:  now.Add(400 * time.Millisecond),
+			EndedAt:    now.Add(500 * time.Millisecond),
+			BytesTx:    30,
+		},
+		{
+			SessionID:  "s1",
+			ResourceID: 1,
+			SourceAddr: "10.0.0.1:5000",
+			DestAddr:   "192.168.1.100:443",
+			Protocol:   "tcp",
+			StartedAt:  now,
+			EndedAt:    now.Add(100 * time.Millisecond),
+			BytesTx:    10,
+		},
+		{
+			SessionID:  "s2",
+			ResourceID: 1,
+			SourceAddr: "10.0.0.1:5001",
+			DestAddr:   "192.168.1.100:443",
+			Protocol:   "tcp",
+			StartedAt:  now.Add(200 * time.Millisecond),
+			EndedAt:    now.Add(300 * time.Millisecond),
+			BytesTx:    20,
+		},
+	}
+
+	result := consolidateSessions(sessions)
+	if len(result) != 1 {
+		t.Fatalf("expected 1 consolidated session, got %d", len(result))
+	}
+
+	s := result[0]
+	if s.ConnectionCount != 3 {
+		t.Errorf("expected ConnectionCount=3, got %d", s.ConnectionCount)
+	}
+	if s.StartedAt != now {
+		t.Errorf("expected StartedAt to be earliest time")
+	}
+	if s.EndedAt != now.Add(500*time.Millisecond) {
+		t.Errorf("expected EndedAt to be latest time")
+	}
+	if s.BytesTx != 60 {
+		t.Errorf("expected BytesTx=60, got %d", s.BytesTx)
+	}
+}
+
+func TestConsolidateSessions_ExactlyAtGapThreshold(t *testing.T) {
+	now := time.Now()
+	sessions := []*AccessSession{
+		{
+			SessionID:  "s1",
+			ResourceID: 1,
+			SourceAddr: "10.0.0.1:5000",
+			DestAddr:   "192.168.1.100:443",
+			Protocol:   "tcp",
+			StartedAt:  now,
+			EndedAt:    now.Add(100 * time.Millisecond),
+		},
+		{
+			// Starts exactly sessionGapThreshold after s1 ends — should still merge
+			SessionID:  "s2",
+			ResourceID: 1,
+			SourceAddr: "10.0.0.1:5001",
+			DestAddr:   "192.168.1.100:443",
+			Protocol:   "tcp",
+			StartedAt:  now.Add(100*time.Millisecond + sessionGapThreshold),
+			EndedAt:    now.Add(100*time.Millisecond + sessionGapThreshold + 50*time.Millisecond),
+		},
+	}
+
+	result := consolidateSessions(sessions)
+	if len(result) != 1 {
+		t.Fatalf("expected 1 session (gap exactly at threshold merges), got %d", len(result))
+	}
+	if result[0].ConnectionCount != 2 {
+		t.Errorf("expected ConnectionCount=2, got %d", result[0].ConnectionCount)
+	}
+}
+
+func TestConsolidateSessions_JustOverGapThreshold(t *testing.T) {
+	now := time.Now()
+	sessions := []*AccessSession{
+		{
+			SessionID:  "s1",
+			ResourceID: 1,
+			SourceAddr: "10.0.0.1:5000",
+			DestAddr:   "192.168.1.100:443",
+			Protocol:   "tcp",
+			StartedAt:  now,
+			EndedAt:    now.Add(100 * time.Millisecond),
+		},
+		{
+			// Starts 1ms over the gap threshold after s1 ends — should split
+			SessionID:  "s2",
+			ResourceID: 1,
+			SourceAddr: "10.0.0.1:5001",
+			DestAddr:   "192.168.1.100:443",
+			Protocol:   "tcp",
+			StartedAt:  now.Add(100*time.Millisecond + sessionGapThreshold + 1*time.Millisecond),
+			EndedAt:    now.Add(100*time.Millisecond + sessionGapThreshold + 50*time.Millisecond),
+		},
+	}
+
+	result := consolidateSessions(sessions)
+	if len(result) != 2 {
+		t.Fatalf("expected 2 sessions (gap just over threshold splits), got %d", len(result))
+	}
+}
+
+func TestConsolidateSessions_UDPSessions(t *testing.T) {
+	now := time.Now()
+	sessions := []*AccessSession{
+		{
+			SessionID:  "u1",
+			ResourceID: 5,
+			SourceAddr: "10.0.0.1:6000",
+			DestAddr:   "192.168.1.100:53",
+			Protocol:   "udp",
+			StartedAt:  now,
+			EndedAt:    now.Add(50 * time.Millisecond),
+			BytesTx:    64,
+			BytesRx:    512,
+		},
+		{
+			SessionID:  "u2",
+			ResourceID: 5,
+			SourceAddr: "10.0.0.1:6001",
+			DestAddr:   "192.168.1.100:53",
+			Protocol:   "udp",
+			StartedAt:  now.Add(100 * time.Millisecond),
+			EndedAt:    now.Add(150 * time.Millisecond),
+			BytesTx:    64,
+			BytesRx:    256,
+		},
+		{
+			SessionID:  "u3",
+			ResourceID: 5,
+			SourceAddr: "10.0.0.1:6002",
+			DestAddr:   "192.168.1.100:53",
+			Protocol:   "udp",
+			StartedAt:  now.Add(200 * time.Millisecond),
+			EndedAt:    now.Add(250 * time.Millisecond),
+			BytesTx:    64,
+			BytesRx:    128,
+		},
+	}
+
+	result := consolidateSessions(sessions)
+	if len(result) != 1 {
+		t.Fatalf("expected 1 consolidated UDP session, got %d", len(result))
+	}
+
+	s := result[0]
+	if s.Protocol != "udp" {
+		t.Errorf("expected protocol=udp, got %q", s.Protocol)
+	}
+	if s.ConnectionCount != 3 {
+		t.Errorf("expected ConnectionCount=3, got %d", s.ConnectionCount)
+	}
+	if s.SourceAddr != "10.0.0.1" {
+		t.Errorf("expected source addr to be IP only, got %q", s.SourceAddr)
+	}
+	if s.BytesTx != 192 {
+		t.Errorf("expected BytesTx=192, got %d", s.BytesTx)
+	}
+	if s.BytesRx != 896 {
+		t.Errorf("expected BytesRx=896, got %d", s.BytesRx)
+	}
+}
+
+func TestConsolidateSessions_MixedGroupsSomeConsolidatedSomeNot(t *testing.T) {
+	now := time.Now()
+	sessions := []*AccessSession{
+		// Group 1: 3 connections to :443 from same IP — should consolidate
+		{
+			SessionID:  "s1",
+			ResourceID: 1,
+			SourceAddr: "10.0.0.1:5000",
+			DestAddr:   "192.168.1.100:443",
+			Protocol:   "tcp",
+			StartedAt:  now,
+			EndedAt:    now.Add(100 * time.Millisecond),
+		},
+		{
+			SessionID:  "s2",
+			ResourceID: 1,
+			SourceAddr: "10.0.0.1:5001",
+			DestAddr:   "192.168.1.100:443",
+			Protocol:   "tcp",
+			StartedAt:  now.Add(200 * time.Millisecond),
+			EndedAt:    now.Add(300 * time.Millisecond),
+		},
+		{
+			SessionID:  "s3",
+			ResourceID: 1,
+			SourceAddr: "10.0.0.1:5002",
+			DestAddr:   "192.168.1.100:443",
+			Protocol:   "tcp",
+			StartedAt:  now.Add(400 * time.Millisecond),
+			EndedAt:    now.Add(500 * time.Millisecond),
+		},
+		// Group 2: 1 connection to :8080 from different IP — should pass through
+		{
+			SessionID:  "s4",
+			ResourceID: 2,
+			SourceAddr: "10.0.0.2:6000",
+			DestAddr:   "192.168.1.200:8080",
+			Protocol:   "tcp",
+			StartedAt:  now.Add(1 * time.Second),
+			EndedAt:    now.Add(2 * time.Second),
+		},
+	}
+
+	result := consolidateSessions(sessions)
+	if len(result) != 2 {
+		t.Fatalf("expected 2 sessions total, got %d", len(result))
+	}
+
+	var consolidated, passthrough *AccessSession
+	for _, s := range result {
+		if s.ConnectionCount > 1 {
+			consolidated = s
+		} else {
+			passthrough = s
+		}
+	}
+
+	if consolidated == nil {
+		t.Fatal("expected a consolidated session")
+	}
+	if consolidated.ConnectionCount != 3 {
+		t.Errorf("consolidated: expected ConnectionCount=3, got %d", consolidated.ConnectionCount)
+	}
+
+	if passthrough == nil {
+		t.Fatal("expected a passthrough session")
+	}
+	if passthrough.SessionID != "s4" {
+		t.Errorf("passthrough: expected session s4, got %s", passthrough.SessionID)
+	}
+}
+
+func TestConsolidateSessions_OverlappingConnections(t *testing.T) {
+	now := time.Now()
+	// Connections that overlap in time (not sequential)
+	sessions := []*AccessSession{
+		{
+			SessionID:  "s1",
+			ResourceID: 1,
+			SourceAddr: "10.0.0.1:5000",
+			DestAddr:   "192.168.1.100:443",
+			Protocol:   "tcp",
+			StartedAt:  now,
+			EndedAt:    now.Add(5 * time.Second),
+			BytesTx:    100,
+		},
+		{
+			SessionID:  "s2",
+			ResourceID: 1,
+			SourceAddr: "10.0.0.1:5001",
+			DestAddr:   "192.168.1.100:443",
+			Protocol:   "tcp",
+			StartedAt:  now.Add(1 * time.Second),
+			EndedAt:    now.Add(3 * time.Second),
+			BytesTx:    200,
+		},
+		{
+			SessionID:  "s3",
+			ResourceID: 1,
+			SourceAddr: "10.0.0.1:5002",
+			DestAddr:   "192.168.1.100:443",
+			Protocol:   "tcp",
+			StartedAt:  now.Add(2 * time.Second),
+			EndedAt:    now.Add(6 * time.Second),
+			BytesTx:    300,
+		},
+	}
+
+	result := consolidateSessions(sessions)
+	if len(result) != 1 {
+		t.Fatalf("expected 1 consolidated session, got %d", len(result))
+	}
+
+	s := result[0]
+	if s.ConnectionCount != 3 {
+		t.Errorf("expected ConnectionCount=3, got %d", s.ConnectionCount)
+	}
+	if s.StartedAt != now {
+		t.Error("expected StartedAt to be earliest")
+	}
+	if s.EndedAt != now.Add(6*time.Second) {
+		t.Error("expected EndedAt to be the latest end time")
+	}
+	if s.BytesTx != 600 {
+		t.Errorf("expected BytesTx=600, got %d", s.BytesTx)
+	}
+}
+
+func TestConsolidateSessions_DoesNotMutateOriginals(t *testing.T) {
+	now := time.Now()
+	s1 := &AccessSession{
+		SessionID:  "s1",
+		ResourceID: 1,
+		SourceAddr: "10.0.0.1:5000",
+		DestAddr:   "192.168.1.100:443",
+		Protocol:   "tcp",
+		StartedAt:  now,
+		EndedAt:    now.Add(100 * time.Millisecond),
+		BytesTx:    100,
+	}
+	s2 := &AccessSession{
+		SessionID:  "s2",
+		ResourceID: 1,
+		SourceAddr: "10.0.0.1:5001",
+		DestAddr:   "192.168.1.100:443",
+		Protocol:   "tcp",
+		StartedAt:  now.Add(200 * time.Millisecond),
+		EndedAt:    now.Add(300 * time.Millisecond),
+		BytesTx:    200,
+	}
+
+	// Save original values
+	origS1Addr := s1.SourceAddr
+	origS1Bytes := s1.BytesTx
+	origS2Addr := s2.SourceAddr
+	origS2Bytes := s2.BytesTx
+
+	_ = consolidateSessions([]*AccessSession{s1, s2})
+
+	if s1.SourceAddr != origS1Addr {
+		t.Errorf("s1.SourceAddr was mutated: %q -> %q", origS1Addr, s1.SourceAddr)
+	}
+	if s1.BytesTx != origS1Bytes {
+		t.Errorf("s1.BytesTx was mutated: %d -> %d", origS1Bytes, s1.BytesTx)
+	}
+	if s2.SourceAddr != origS2Addr {
+		t.Errorf("s2.SourceAddr was mutated: %q -> %q", origS2Addr, s2.SourceAddr)
+	}
+	if s2.BytesTx != origS2Bytes {
+		t.Errorf("s2.BytesTx was mutated: %d -> %d", origS2Bytes, s2.BytesTx)
+	}
+}
+
+func TestConsolidateSessions_ThreeBurstsWithGaps(t *testing.T) {
+	now := time.Now()
+
+	sessions := make([]*AccessSession, 0, 9)
+
+	// Burst 1: 3 connections at t=0
+	for i := 0; i < 3; i++ {
+		sessions = append(sessions, &AccessSession{
+			SessionID:  generateSessionID(),
+			ResourceID: 1,
+			SourceAddr: "10.0.0.1:" + string(rune('A'+i)),
+			DestAddr:   "192.168.1.100:443",
+			Protocol:   "tcp",
+			StartedAt:  now.Add(time.Duration(i*100) * time.Millisecond),
+			EndedAt:    now.Add(time.Duration(i*100+50) * time.Millisecond),
+		})
+	}
+
+	// Burst 2: 3 connections at t=20s (well past the 5s gap)
+	for i := 0; i < 3; i++ {
+		sessions = append(sessions, &AccessSession{
+			SessionID:  generateSessionID(),
+			ResourceID: 1,
+			SourceAddr: "10.0.0.1:" + string(rune('D'+i)),
+			DestAddr:   "192.168.1.100:443",
+			Protocol:   "tcp",
+			StartedAt:  now.Add(20*time.Second + time.Duration(i*100)*time.Millisecond),
+			EndedAt:    now.Add(20*time.Second + time.Duration(i*100+50)*time.Millisecond),
+		})
+	}
+
+	// Burst 3: 3 connections at t=40s
+	for i := 0; i < 3; i++ {
+		sessions = append(sessions, &AccessSession{
+			SessionID:  generateSessionID(),
+			ResourceID: 1,
+			SourceAddr: "10.0.0.1:" + string(rune('G'+i)),
+			DestAddr:   "192.168.1.100:443",
+			Protocol:   "tcp",
+			StartedAt:  now.Add(40*time.Second + time.Duration(i*100)*time.Millisecond),
+			EndedAt:    now.Add(40*time.Second + time.Duration(i*100+50)*time.Millisecond),
+		})
+	}
+
+	result := consolidateSessions(sessions)
+	if len(result) != 3 {
+		t.Fatalf("expected 3 consolidated sessions (3 bursts), got %d", len(result))
+	}
+
+	for _, s := range result {
+		if s.ConnectionCount != 3 {
+			t.Errorf("expected each burst to have ConnectionCount=3, got %d (started=%v)", s.ConnectionCount, s.StartedAt)
+		}
+	}
+}
+
+func TestFinalizeMergedSourceAddr(t *testing.T) {
+	s := &AccessSession{SourceAddr: "10.0.0.1:5000"}
+	ports := map[string]struct{}{"10.0.0.1:5000": {}}
+	finalizeMergedSourceAddr(s, "10.0.0.1", ports)
+	if s.SourceAddr != "10.0.0.1:5000" {
+		t.Errorf("single port: expected addr preserved, got %q", s.SourceAddr)
+	}
+
+	s2 := &AccessSession{SourceAddr: "10.0.0.1:5000"}
+	ports2 := map[string]struct{}{"10.0.0.1:5000": {}, "10.0.0.1:5001": {}}
+	finalizeMergedSourceAddr(s2, "10.0.0.1", ports2)
+	if s2.SourceAddr != "10.0.0.1" {
+		t.Errorf("multiple ports: expected IP only, got %q", s2.SourceAddr)
+	}
+}
+
+func TestCloneSession(t *testing.T) {
+	original := &AccessSession{
+		SessionID:  "test",
+		ResourceID: 42,
+		SourceAddr: "1.2.3.4:100",
+		DestAddr:   "5.6.7.8:443",
+		Protocol:   "tcp",
+		BytesTx:    999,
+	}
+
+	clone := cloneSession(original)
+
+	if clone == original {
+		t.Error("clone should be a different pointer")
+	}
+	if clone.SessionID != original.SessionID {
+		t.Error("clone should have same SessionID")
+	}
+
+	// Mutating clone should not affect original
+	clone.BytesTx = 0
+	clone.SourceAddr = "changed"
+	if original.BytesTx != 999 {
+		t.Error("mutating clone affected original BytesTx")
+	}
+	if original.SourceAddr != "1.2.3.4:100" {
+		t.Error("mutating clone affected original SourceAddr")
+	}
+}

netstack2/handlers.go

@@ -137,14 +137,31 @@ func (h *TCPHandler) InstallTCPHandler() error {
 
 // handleTCPConn handles a TCP connection by proxying it to the actual target
 func (h *TCPHandler) handleTCPConn(netstackConn *gonet.TCPConn, id stack.TransportEndpointID) {
-	defer netstackConn.Close()
-
-	// Extract source and target address from the connection ID
+	// Extract source and target address from the connection ID first so they
+	// are available for HTTP routing before any defer is set up.
 	srcIP := id.RemoteAddress.String()
 	srcPort := id.RemotePort
 	dstIP := id.LocalAddress.String()
 	dstPort := id.LocalPort
 
+	// For HTTP/HTTPS ports, look up the matching subnet rule. If the rule has
+	// Protocol configured, hand the connection off to the HTTP handler which
+	// takes full ownership of the lifecycle (the defer close must not be
+	// installed before this point).
+	if (dstPort == 80 || dstPort == 443) && h.proxyHandler != nil && h.proxyHandler.httpHandler != nil {
+		srcAddr, _ := netip.ParseAddr(srcIP)
+		dstAddr, _ := netip.ParseAddr(dstIP)
+		rule := h.proxyHandler.subnetLookup.Match(srcAddr, dstAddr, dstPort, tcp.ProtocolNumber)
+		if rule != nil && rule.Protocol != "" {
+			logger.Info("TCP Forwarder: Routing %s:%d -> %s:%d to HTTP handler (%s)",
+				srcIP, srcPort, dstIP, dstPort, rule.Protocol)
+			h.proxyHandler.httpHandler.HandleConn(netstackConn, rule)
+			return
+		}
+	}
+
+	defer netstackConn.Close()
+
 	logger.Info("TCP Forwarder: Handling connection %s:%d -> %s:%d", srcIP, srcPort, dstIP, dstPort)
 
 	// Check if there's a destination rewrite for this connection (e.g., localhost targets)
@@ -158,6 +175,18 @@ func (h *TCPHandler) handleTCPConn(netstackConn *gonet.TCPConn, id stack.Transpo
 
 	targetAddr := fmt.Sprintf("%s:%d", actualDstIP, dstPort)
 
+	// Look up resource ID and start access session if applicable
+	var accessSessionID string
+	if h.proxyHandler != nil {
+		resourceId := h.proxyHandler.LookupResourceId(srcIP, dstIP, dstPort, uint8(tcp.ProtocolNumber))
+		if resourceId != 0 {
+			if al := h.proxyHandler.GetAccessLogger(); al != nil {
+				srcAddr := fmt.Sprintf("%s:%d", srcIP, srcPort)
+				accessSessionID = al.StartTCPSession(resourceId, srcAddr, targetAddr)
+			}
+		}
+	}
+
 	// Create context with timeout for connection establishment
 	ctx, cancel := context.WithTimeout(context.Background(), tcpConnectTimeout)
 	defer cancel()
@@ -167,11 +196,26 @@ func (h *TCPHandler) handleTCPConn(netstackConn *gonet.TCPConn, id stack.Transpo
 	targetConn, err := d.DialContext(ctx, "tcp", targetAddr)
 	if err != nil {
 		logger.Info("TCP Forwarder: Failed to connect to %s: %v", targetAddr, err)
+		// End access session on connection failure
+		if accessSessionID != "" {
+			if al := h.proxyHandler.GetAccessLogger(); al != nil {
+				al.EndTCPSession(accessSessionID)
+			}
+		}
 		// Connection failed, netstack will handle RST
 		return
 	}
 	defer targetConn.Close()
 
+	// End access session when connection closes
+	if accessSessionID != "" {
+		defer func() {
+			if al := h.proxyHandler.GetAccessLogger(); al != nil {
+				al.EndTCPSession(accessSessionID)
+			}
+		}()
+	}
+
 	logger.Info("TCP Forwarder: Successfully connected to %s, starting bidirectional copy", targetAddr)
 
 	// Bidirectional copy between netstack and target
@@ -280,6 +324,27 @@ func (h *UDPHandler) handleUDPConn(netstackConn *gonet.UDPConn, id stack.Transpo
 
 	targetAddr := fmt.Sprintf("%s:%d", actualDstIP, dstPort)
 
+	// Look up resource ID and start access session if applicable
+	var accessSessionID string
+	if h.proxyHandler != nil {
+		resourceId := h.proxyHandler.LookupResourceId(srcIP, dstIP, dstPort, uint8(udp.ProtocolNumber))
+		if resourceId != 0 {
+			if al := h.proxyHandler.GetAccessLogger(); al != nil {
+				srcAddr := fmt.Sprintf("%s:%d", srcIP, srcPort)
+				accessSessionID = al.TrackUDPSession(resourceId, srcAddr, targetAddr)
+			}
+		}
+	}
+
+	// End access session when UDP handler returns (timeout or error)
+	if accessSessionID != "" {
+		defer func() {
+			if al := h.proxyHandler.GetAccessLogger(); al != nil {
+				al.EndUDPSession(accessSessionID)
+			}
+		}()
+	}
+
 	// Resolve target address
 	remoteUDPAddr, err := net.ResolveUDPAddr("udp", targetAddr)
 	if err != nil {

netstack2/http_handler.go

@@ -0,0 +1,318 @@
+/* SPDX-License-Identifier: MIT
+ *
+ * Copyright (C) 2017-2025 WireGuard LLC. All Rights Reserved.
+ */
+
+package netstack2
+
+import (
+	"context"
+	"crypto/tls"
+	"fmt"
+	"net"
+	"net/http"
+	"net/http/httputil"
+	"net/url"
+	"sync"
+
+	"github.com/fosrl/newt/logger"
+	"gvisor.dev/gvisor/pkg/tcpip/stack"
+)
+
+// ---------------------------------------------------------------------------
+// HTTPTarget
+// ---------------------------------------------------------------------------
+
+// HTTPTarget describes a single downstream HTTP or HTTPS service that the
+// proxy should forward requests to.
+type HTTPTarget struct {
+	DestAddr string `json:"destAddr"` // IP address or hostname of the downstream service
+	DestPort uint16 `json:"destPort"` // TCP port of the downstream service
+	UseHTTPS bool   `json:"useHttps"` // When true the outbound leg uses HTTPS
+}
+
+// ---------------------------------------------------------------------------
+// HTTPHandler
+// ---------------------------------------------------------------------------
+
+// HTTPHandler intercepts TCP connections from the netstack forwarder on ports
+// 80 and 443 and services them as HTTP or HTTPS, reverse-proxying each request
+// to downstream targets specified by the matching SubnetRule.
+//
+// HTTP and raw TCP are fully separate: a connection is only routed here when
+// its SubnetRule has Protocol set ("http" or "https"). All other connections
+// on those ports fall through to the normal raw-TCP path.
+//
+// Incoming TLS termination (Protocol == "https") is performed per-connection
+// using the certificate and key stored in the rule, so different subnet rules
+// can present different certificates without sharing any state.
+//
+// Outbound connections to downstream targets honour HTTPTarget.UseHTTPS
+// independently of the incoming protocol.
+type HTTPHandler struct {
+	stack        *stack.Stack
+	proxyHandler *ProxyHandler
+
+	listener *chanListener
+	server   *http.Server
+
+	// proxyCache holds pre-built *httputil.ReverseProxy values keyed by the
+	// canonical target URL string ("scheme://host:port"). Building a proxy is
+	// cheap, but reusing one preserves the underlying http.Transport connection
+	// pool, which matters for throughput.
+	proxyCache sync.Map // map[string]*httputil.ReverseProxy
+
+	// tlsCache holds pre-parsed *tls.Config values keyed by the concatenation
+	// of the PEM certificate and key. Parsing a keypair is relatively expensive
+	// and the same cert is likely reused across many connections.
+	tlsCache sync.Map // map[string]*tls.Config
+}
+
+// ---------------------------------------------------------------------------
+// chanListener – net.Listener backed by a channel
+// ---------------------------------------------------------------------------
+
+// chanListener implements net.Listener by receiving net.Conn values over a
+// buffered channel. This lets the netstack TCP forwarder hand off connections
+// directly to a running http.Server without any real OS socket.
+type chanListener struct {
+	connCh chan net.Conn
+	closed chan struct{}
+	once   sync.Once
+}
+
+func newChanListener() *chanListener {
+	return &chanListener{
+		connCh: make(chan net.Conn, 128),
+		closed: make(chan struct{}),
+	}
+}
+
+// Accept blocks until a connection is available or the listener is closed.
+func (l *chanListener) Accept() (net.Conn, error) {
+	select {
+	case conn, ok := <-l.connCh:
+		if !ok {
+			return nil, net.ErrClosed
+		}
+		return conn, nil
+	case <-l.closed:
+		return nil, net.ErrClosed
+	}
+}
+
+// Close shuts down the listener; subsequent Accept calls return net.ErrClosed.
+func (l *chanListener) Close() error {
+	l.once.Do(func() { close(l.closed) })
+	return nil
+}
+
+// Addr returns a placeholder address (the listener has no real OS socket).
+func (l *chanListener) Addr() net.Addr {
+	return &net.TCPAddr{}
+}
+
+// send delivers conn to the listener. Returns false if the listener is already
+// closed, in which case the caller is responsible for closing conn.
+func (l *chanListener) send(conn net.Conn) bool {
+	select {
+	case l.connCh <- conn:
+		return true
+	case <-l.closed:
+		return false
+	}
+}
+
+// ---------------------------------------------------------------------------
+// httpConnCtx – conn wrapper that carries a SubnetRule through the listener
+// ---------------------------------------------------------------------------
+
+// httpConnCtx wraps a net.Conn so the matching SubnetRule can be passed
+// through the chanListener into the http.Server's ConnContext callback,
+// making it available to request handlers via the request context.
+type httpConnCtx struct {
+	net.Conn
+	rule *SubnetRule
+}
+
+// connCtxKey is the unexported context key used to store a *SubnetRule on the
+// per-connection context created by http.Server.ConnContext.
+type connCtxKey struct{}
+
+// ---------------------------------------------------------------------------
+// Constructor and lifecycle
+// ---------------------------------------------------------------------------
+
+// NewHTTPHandler creates an HTTPHandler attached to the given stack and
+// ProxyHandler. Call Start to begin serving connections.
+func NewHTTPHandler(s *stack.Stack, ph *ProxyHandler) *HTTPHandler {
+	return &HTTPHandler{
+		stack:        s,
+		proxyHandler: ph,
+	}
+}
+
+// Start launches the internal http.Server that services connections delivered
+// via HandleConn. The server runs for the lifetime of the HTTPHandler; call
+// Close to stop it.
+func (h *HTTPHandler) Start() error {
+	h.listener = newChanListener()
+
+	h.server = &http.Server{
+		Handler: http.HandlerFunc(h.handleRequest),
+		// ConnContext runs once per accepted connection and attaches the
+		// SubnetRule carried by httpConnCtx to the connection's context so
+		// that handleRequest can retrieve it without any global state.
+		ConnContext: func(ctx context.Context, c net.Conn) context.Context {
+			if cc, ok := c.(*httpConnCtx); ok {
+				return context.WithValue(ctx, connCtxKey{}, cc.rule)
+			}
+			return ctx
+		},
+	}
+
+	go func() {
+		if err := h.server.Serve(h.listener); err != nil && err != http.ErrServerClosed {
+			logger.Error("HTTP handler: server exited unexpectedly: %v", err)
+		}
+	}()
+
+	logger.Info("HTTP handler: ready — routing determined per SubnetRule on ports 80/443")
+	return nil
+}
+
+// HandleConn accepts a TCP connection from the netstack forwarder together
+// with the SubnetRule that matched it. The HTTP handler takes full ownership
+// of the connection's lifecycle; the caller must NOT close conn after this call.
+//
+// When rule.Protocol is "https", TLS termination is performed on conn using
+// the certificate and key stored in rule.TLSCert and rule.TLSKey before the
+// connection is passed to the HTTP server. The HTTP server itself is always
+// plain-HTTP; TLS is fully unwrapped at this layer.
+func (h *HTTPHandler) HandleConn(conn net.Conn, rule *SubnetRule) {
+	var effectiveConn net.Conn = conn
+
+	if rule.Protocol == "https" {
+		tlsCfg, err := h.getTLSConfig(rule)
+		if err != nil {
+			logger.Error("HTTP handler: cannot build TLS config for connection from %s: %v",
+				conn.RemoteAddr(), err)
+			conn.Close()
+			return
+		}
+		// tls.Server wraps the raw conn; the TLS handshake is deferred until
+		// the first Read, which the http.Server will trigger naturally.
+		effectiveConn = tls.Server(conn, tlsCfg)
+	}
+
+	wrapped := &httpConnCtx{Conn: effectiveConn, rule: rule}
+	if !h.listener.send(wrapped) {
+		// Listener is already closed — clean up the orphaned connection.
+		effectiveConn.Close()
+	}
+}
+
+// Close gracefully shuts down the HTTP server and the underlying channel
+// listener, causing the goroutine started in Start to exit.
+func (h *HTTPHandler) Close() error {
+	if h.server != nil {
+		if err := h.server.Close(); err != nil {
+			return err
+		}
+	}
+	if h.listener != nil {
+		h.listener.Close()
+	}
+	return nil
+}
+
+// ---------------------------------------------------------------------------
+// Internal helpers
+// ---------------------------------------------------------------------------
+
+// getTLSConfig returns a *tls.Config for the cert/key pair in rule, using a
+// cache to avoid re-parsing the same keypair on every connection.
+// The cache key is the concatenation of the PEM cert and key strings, so
+// different rules that happen to share the same material hit the same entry.
+func (h *HTTPHandler) getTLSConfig(rule *SubnetRule) (*tls.Config, error) {
+	cacheKey := rule.TLSCert + "|" + rule.TLSKey
+	if v, ok := h.tlsCache.Load(cacheKey); ok {
+		return v.(*tls.Config), nil
+	}
+
+	cert, err := tls.X509KeyPair([]byte(rule.TLSCert), []byte(rule.TLSKey))
+	if err != nil {
+		return nil, fmt.Errorf("failed to parse TLS keypair: %w", err)
+	}
+	cfg := &tls.Config{
+		Certificates: []tls.Certificate{cert},
+	}
+	// LoadOrStore is safe under concurrent calls: if two goroutines race here
+	// both will produce a valid config; the loser's work is discarded.
+	actual, _ := h.tlsCache.LoadOrStore(cacheKey, cfg)
+	return actual.(*tls.Config), nil
+}
+
+// getProxy returns a cached *httputil.ReverseProxy for the given target,
+// creating one on first use. Reusing the proxy preserves its http.Transport
+// connection pool, avoiding repeated TCP/TLS handshakes to the downstream.
+func (h *HTTPHandler) getProxy(target HTTPTarget) *httputil.ReverseProxy {
+	scheme := "http"
+	if target.UseHTTPS {
+		scheme = "https"
+	}
+	cacheKey := fmt.Sprintf("%s://%s:%d", scheme, target.DestAddr, target.DestPort)
+
+	if v, ok := h.proxyCache.Load(cacheKey); ok {
+		return v.(*httputil.ReverseProxy)
+	}
+
+	targetURL := &url.URL{
+		Scheme: scheme,
+		Host:   fmt.Sprintf("%s:%d", target.DestAddr, target.DestPort),
+	}
+	proxy := httputil.NewSingleHostReverseProxy(targetURL)
+
+	if target.UseHTTPS {
+		// Allow self-signed certificates on downstream HTTPS targets.
+		proxy.Transport = &http.Transport{
+			TLSClientConfig: &tls.Config{
+				InsecureSkipVerify: true, //nolint:gosec // downstream self-signed certs are a supported configuration
+			},
+		}
+	}
+
+	proxy.ErrorHandler = func(w http.ResponseWriter, r *http.Request, err error) {
+		logger.Error("HTTP handler: upstream error (%s %s -> %s): %v",
+			r.Method, r.URL.RequestURI(), cacheKey, err)
+		http.Error(w, "Bad Gateway", http.StatusBadGateway)
+	}
+
+	actual, _ := h.proxyCache.LoadOrStore(cacheKey, proxy)
+	return actual.(*httputil.ReverseProxy)
+}
+
+// handleRequest is the http.Handler entry point. It retrieves the SubnetRule
+// attached to the connection by ConnContext, selects the first configured
+// downstream target, and forwards the request via the cached ReverseProxy.
+//
+// TODO: add host/path-based routing across multiple HTTPTargets once the
+// configuration model evolves beyond a single target per rule.
+func (h *HTTPHandler) handleRequest(w http.ResponseWriter, r *http.Request) {
+	rule, _ := r.Context().Value(connCtxKey{}).(*SubnetRule)
+	if rule == nil || len(rule.HTTPTargets) == 0 {
+		logger.Error("HTTP handler: no downstream targets for request %s %s", r.Method, r.URL.RequestURI())
+		http.Error(w, "no targets configured", http.StatusBadGateway)
+		return
+	}
+
+	target := rule.HTTPTargets[0]
+	scheme := "http"
+	if target.UseHTTPS {
+		scheme = "https"
+	}
+	logger.Info("HTTP handler: %s %s -> %s://%s:%d",
+		r.Method, r.URL.RequestURI(), scheme, target.DestAddr, target.DestPort)
+
+	h.getProxy(target).ServeHTTP(w, r)
+}

netstack2/proxy.go

@@ -22,6 +22,12 @@ import (
 	"gvisor.dev/gvisor/pkg/tcpip/transport/udp"
 )
 
+const (
+	// udpAccessSessionTimeout is how long a UDP access session stays alive without traffic
+	// before being considered ended by the access logger
+	udpAccessSessionTimeout = 120 * time.Second
+)
+
 // PortRange represents an allowed range of ports (inclusive) with optional protocol filtering
 // Protocol can be "tcp", "udp", or "" (empty string means both protocols)
 type PortRange struct {
@@ -46,6 +52,15 @@ type SubnetRule struct {
 	DisableIcmp  bool         // If true, ICMP traffic is blocked for this subnet
 	RewriteTo    string       // Optional rewrite address for DNAT - can be IP/CIDR or domain name
 	PortRanges   []PortRange  // empty slice means all ports allowed
+	ResourceId   int          // Optional resource ID from the server for access logging
+
+	// HTTP proxy configuration (optional).
+	// When Protocol is non-empty the TCP connection is handled by HTTPHandler
+	// instead of the raw TCP forwarder.
+	Protocol    string       // "", "http", or "https" — controls the incoming (client-facing) protocol
+	HTTPTargets []HTTPTarget // downstream services to proxy requests to
+	TLSCert     string       // PEM-encoded certificate for incoming HTTPS termination
+	TLSKey      string       // PEM-encoded private key for incoming HTTPS termination
 }
 
 // GetAllRules returns a copy of all subnet rules
@@ -107,14 +122,17 @@ type ProxyHandler struct {
 	tcpHandler        *TCPHandler
 	udpHandler        *UDPHandler
 	icmpHandler       *ICMPHandler
+	httpHandler       *HTTPHandler
 	subnetLookup      *SubnetLookup
 	natTable          map[connKey]*natState
 	reverseNatTable   map[reverseConnKey]*natState // Reverse lookup map for O(1) reply packet NAT
 	destRewriteTable  map[destKey]netip.Addr       // Maps original dest to rewritten dest for handler lookups
+	resourceTable     map[destKey]int              // Maps connection key to resource ID for access logging
 	natMu             sync.RWMutex
 	enabled           bool
 	icmpReplies       chan []byte          // Channel for ICMP reply packets to be sent back through the tunnel
 	notifiable        channel.Notification // Notification handler for triggering reads
+	accessLogger      *AccessLogger        // Access logger for tracking sessions
 }
 
 // ProxyHandlerOptions configures the proxy handler
@@ -137,7 +155,9 @@ func NewProxyHandler(options ProxyHandlerOptions) (*ProxyHandler, error) {
 		natTable:         make(map[connKey]*natState),
 		reverseNatTable:  make(map[reverseConnKey]*natState),
 		destRewriteTable: make(map[destKey]netip.Addr),
+		resourceTable:    make(map[destKey]int),
 		icmpReplies:      make(chan []byte, 256), // Buffer for ICMP reply packets
+		accessLogger:     NewAccessLogger(udpAccessSessionTimeout),
 		proxyEp:          channel.New(1024, uint32(options.MTU), ""),
 		proxyStack: stack.New(stack.Options{
 			NetworkProtocols: []stack.NetworkProtocolFactory{
@@ -153,12 +173,21 @@ func NewProxyHandler(options ProxyHandlerOptions) (*ProxyHandler, error) {
 		}),
 	}
 
-	// Initialize TCP handler if enabled
+	// Initialize TCP handler if enabled. The HTTP handler piggybacks on the
+	// TCP forwarder — TCPHandler.handleTCPConn checks the subnet rule for
+	// ports 80/443 and routes matching connections to the HTTP handler, so
+	// the HTTP handler is always initialised alongside TCP.
 	if options.EnableTCP {
 		handler.tcpHandler = NewTCPHandler(handler.proxyStack, handler)
 		if err := handler.tcpHandler.InstallTCPHandler(); err != nil {
 			return nil, fmt.Errorf("failed to install TCP handler: %v", err)
 		}
+
+		handler.httpHandler = NewHTTPHandler(handler.proxyStack, handler)
+		if err := handler.httpHandler.Start(); err != nil {
+			return nil, fmt.Errorf("failed to start HTTP handler: %v", err)
+		}
+		logger.Debug("ProxyHandler: HTTP handler enabled")
 	}
 
 	// Initialize UDP handler if enabled
@@ -197,16 +226,14 @@ func NewProxyHandler(options ProxyHandlerOptions) (*ProxyHandler, error) {
 	return handler, nil
 }
 
-// AddSubnetRule adds a subnet with optional port restrictions to the proxy handler
-// sourcePrefix: The IP prefix of the peer sending the data
-// destPrefix: The IP prefix of the destination
-// rewriteTo: Optional address to rewrite destination to - can be IP/CIDR or domain name
-// If portRanges is nil or empty, all ports are allowed for this subnet
-func (p *ProxyHandler) AddSubnetRule(sourcePrefix, destPrefix netip.Prefix, rewriteTo string, portRanges []PortRange, disableIcmp bool) {
+// AddSubnetRule adds a subnet rule to the proxy handler.
+// HTTP proxy behaviour is configured via rule.Protocol, rule.HTTPTargets,
+// rule.TLSCert, and rule.TLSKey; leave Protocol empty for raw TCP/UDP.
+func (p *ProxyHandler) AddSubnetRule(rule SubnetRule) {
 	if p == nil || !p.enabled {
 		return
 	}
-	p.subnetLookup.AddSubnet(sourcePrefix, destPrefix, rewriteTo, portRanges, disableIcmp)
+	p.subnetLookup.AddSubnet(rule)
 }
 
 // RemoveSubnetRule removes a subnet from the proxy handler
@@ -225,6 +252,43 @@ func (p *ProxyHandler) GetAllRules() []SubnetRule {
 	return p.subnetLookup.GetAllRules()
 }
 
+// LookupResourceId looks up the resource ID for a connection
+// Returns 0 if no resource ID is associated with this connection
+func (p *ProxyHandler) LookupResourceId(srcIP, dstIP string, dstPort uint16, proto uint8) int {
+	if p == nil || !p.enabled {
+		return 0
+	}
+
+	key := destKey{
+		srcIP:   srcIP,
+		dstIP:   dstIP,
+		dstPort: dstPort,
+		proto:   proto,
+	}
+
+	p.natMu.RLock()
+	defer p.natMu.RUnlock()
+
+	return p.resourceTable[key]
+}
+
+// GetAccessLogger returns the access logger for session tracking
+func (p *ProxyHandler) GetAccessLogger() *AccessLogger {
+	if p == nil {
+		return nil
+	}
+	return p.accessLogger
+}
+
+// SetAccessLogSender configures the function used to send compressed access log
+// batches to the server. This should be called once the websocket client is available.
+func (p *ProxyHandler) SetAccessLogSender(fn SendFunc) {
+	if p == nil || !p.enabled || p.accessLogger == nil {
+		return
+	}
+	p.accessLogger.SetSendFunc(fn)
+}
+
 // LookupDestinationRewrite looks up the rewritten destination for a connection
 // This is used by TCP/UDP handlers to find the actual target address
 func (p *ProxyHandler) LookupDestinationRewrite(srcIP, dstIP string, dstPort uint16, proto uint8) (netip.Addr, bool) {
@@ -387,8 +451,22 @@ func (p *ProxyHandler) HandleIncomingPacket(packet []byte) bool {
 	// Check if the source IP, destination IP, port, and protocol match any subnet rule
 	matchedRule := p.subnetLookup.Match(srcAddr, dstAddr, dstPort, protocol)
 	if matchedRule != nil {
-		logger.Debug("HandleIncomingPacket: Matched rule for %s -> %s (proto=%d, port=%d)",
-			srcAddr, dstAddr, protocol, dstPort)
+		logger.Debug("HandleIncomingPacket: Matched rule for %s -> %s (proto=%d, port=%d, resourceId=%d)",
+			srcAddr, dstAddr, protocol, dstPort, matchedRule.ResourceId)
+
+		// Store resource ID for connections without DNAT as well
+		if matchedRule.ResourceId != 0 && matchedRule.RewriteTo == "" {
+			dKey := destKey{
+				srcIP:   srcAddr.String(),
+				dstIP:   dstAddr.String(),
+				dstPort: dstPort,
+				proto:   uint8(protocol),
+			}
+			p.natMu.Lock()
+			p.resourceTable[dKey] = matchedRule.ResourceId
+			p.natMu.Unlock()
+		}
+
 		// Check if we need to perform DNAT
 		if matchedRule.RewriteTo != "" {
 			// Create connection tracking key using original destination
@@ -420,6 +498,13 @@ func (p *ProxyHandler) HandleIncomingPacket(packet []byte) bool {
 				proto:   uint8(protocol),
 			}
 
+			// Store resource ID for access logging if present
+			if matchedRule.ResourceId != 0 {
+				p.natMu.Lock()
+				p.resourceTable[dKey] = matchedRule.ResourceId
+				p.natMu.Unlock()
+			}
+
 			// Check if we already have a NAT entry for this connection
 			p.natMu.RLock()
 			existingEntry, exists := p.natTable[key]
@@ -720,6 +805,16 @@ func (p *ProxyHandler) Close() error {
 		return nil
 	}
 
+	// Shut down access logger
+	if p.accessLogger != nil {
+		p.accessLogger.Close()
+	}
+
+	// Shut down HTTP handler
+	if p.httpHandler != nil {
+		p.httpHandler.Close()
+	}
+
 	// Close ICMP replies channel
 	if p.icmpReplies != nil {
 		close(p.icmpReplies)

netstack2/subnet_lookup.go

@@ -44,23 +44,18 @@ func prefixEqual(a, b netip.Prefix) bool {
 	return a.Masked() == b.Masked()
 }
 
-// AddSubnet adds a subnet rule with source and destination prefixes and optional port restrictions
-// If portRanges is nil or empty, all ports are allowed for this subnet
-// rewriteTo can be either an IP/CIDR (e.g., "192.168.1.1/32") or a domain name (e.g., "example.com")
-func (sl *SubnetLookup) AddSubnet(sourcePrefix, destPrefix netip.Prefix, rewriteTo string, portRanges []PortRange, disableIcmp bool) {
+// AddSubnet adds a subnet rule to the lookup table.
+// If rule.PortRanges is nil or empty, all ports are allowed.
+// rule.RewriteTo can be either an IP/CIDR (e.g., "192.168.1.1/32") or a domain name (e.g., "example.com").
+// HTTP proxy behaviour is driven by rule.Protocol, rule.HTTPTargets, rule.TLSCert, and rule.TLSKey.
+func (sl *SubnetLookup) AddSubnet(rule SubnetRule) {
 	sl.mu.Lock()
 	defer sl.mu.Unlock()
 
-	rule := &SubnetRule{
-		SourcePrefix: sourcePrefix,
-		DestPrefix:   destPrefix,
-		DisableIcmp:  disableIcmp,
-		RewriteTo:    rewriteTo,
-		PortRanges:   portRanges,
-	}
+	rulePtr := &rule
 
 	// Canonicalize source prefix to handle host bits correctly
-	canonicalSourcePrefix := sourcePrefix.Masked()
+	canonicalSourcePrefix := rule.SourcePrefix.Masked()
 
 	// Get or create destination trie for this source prefix
 	destTriePtr, exists := sl.sourceTrie.Get(canonicalSourcePrefix)
@@ -75,12 +70,12 @@ func (sl *SubnetLookup) AddSubnet(sourcePrefix, destPrefix netip.Prefix, rewrite
 
 	// Canonicalize destination prefix to handle host bits correctly
 	// BART masks prefixes internally, so we need to match that behavior in our bookkeeping
-	canonicalDestPrefix := destPrefix.Masked()
+	canonicalDestPrefix := rule.DestPrefix.Masked()
 
 	// Add rule to destination trie
 	// Original behavior: overwrite if same (sourcePrefix, destPrefix) exists
 	// Store as single-element slice to match original overwrite behavior
-	destTriePtr.trie.Insert(canonicalDestPrefix, []*SubnetRule{rule})
+	destTriePtr.trie.Insert(canonicalDestPrefix, []*SubnetRule{rulePtr})
 
 	// Update destTriePtr.rules - remove old rule with same canonical prefix if exists, then add new one
 	// Use canonical comparison to handle cases like 10.0.0.5/24 vs 10.0.0.0/24
@@ -90,7 +85,7 @@ func (sl *SubnetLookup) AddSubnet(sourcePrefix, destPrefix netip.Prefix, rewrite
 			newRules = append(newRules, r)
 		}
 	}
-	newRules = append(newRules, rule)
+	newRules = append(newRules, rulePtr)
 	destTriePtr.rules = newRules
 }
 

netstack2/tun.go

@@ -351,13 +351,13 @@ func (net *Net) ListenUDP(laddr *net.UDPAddr) (*gonet.UDPConn, error) {
 	return net.DialUDP(laddr, nil)
 }
 
-// AddProxySubnetRule adds a subnet rule to the proxy handler
-// If portRanges is nil or empty, all ports are allowed for this subnet
-// rewriteTo can be either an IP/CIDR (e.g., "192.168.1.1/32") or a domain name (e.g., "example.com")
-func (net *Net) AddProxySubnetRule(sourcePrefix, destPrefix netip.Prefix, rewriteTo string, portRanges []PortRange, disableIcmp bool) {
+// AddProxySubnetRule adds a subnet rule to the proxy handler.
+// HTTP proxy behaviour is configured via rule.Protocol, rule.HTTPTargets,
+// rule.TLSCert, and rule.TLSKey; leave Protocol empty for raw TCP/UDP.
+func (net *Net) AddProxySubnetRule(rule SubnetRule) {
 	tun := (*netTun)(net)
 	if tun.proxyHandler != nil {
-		tun.proxyHandler.AddSubnetRule(sourcePrefix, destPrefix, rewriteTo, portRanges, disableIcmp)
+		tun.proxyHandler.AddSubnetRule(rule)
 	}
 }
 
@@ -385,6 +385,15 @@ func (net *Net) GetProxyHandler() *ProxyHandler {
 	return tun.proxyHandler
 }
 
+// SetAccessLogSender configures the function used to send compressed access log
+// batches to the server. This should be called once the websocket client is available.
+func (net *Net) SetAccessLogSender(fn SendFunc) {
+	tun := (*netTun)(net)
+	if tun.proxyHandler != nil {
+		tun.proxyHandler.SetAccessLogSender(fn)
+	}
+}
+
 type PingConn struct {
 	laddr    PingAddr
 	raddr    PingAddr

websocket/client.go

@@ -42,6 +42,7 @@ type Client struct {
 	onTokenUpdate     func(token string)
 	writeMux          sync.Mutex
 	clientType        string // Type of client (e.g., "newt", "olm")
+	configFilePath    string // Optional override for the config file path
 	tlsConfig         TLSConfig
 	metricsCtxMu      sync.RWMutex
 	metricsCtx        context.Context
@@ -52,6 +53,7 @@ type Client struct {
 	processingMessage bool           // Flag to track if a message is currently being processed
 	processingMux     sync.RWMutex   // Protects processingMessage
 	processingWg      sync.WaitGroup // WaitGroup to wait for message processing to complete
+	justProvisioned   bool           // Set to true when provisionIfNeeded exchanges a key for permanent credentials
 }
 
 type ClientOption func(*Client)
@@ -77,6 +79,12 @@ func WithBaseURL(url string) ClientOption {
 }
 
 // WithTLSConfig sets the TLS configuration for the client
+func WithConfigFile(path string) ClientOption {
+	return func(c *Client) {
+		c.configFilePath = path
+	}
+}
+
 func WithTLSConfig(config TLSConfig) ClientOption {
 	return func(c *Client) {
 		c.tlsConfig = config
@@ -95,6 +103,16 @@ func (c *Client) OnTokenUpdate(callback func(token string)) {
 	c.onTokenUpdate = callback
 }
 
+// WasJustProvisioned reports whether the client exchanged a provisioning key
+// for permanent credentials during the most recent connection attempt. It
+// consumes the flag – subsequent calls return false until provisioning occurs
+// again (which, in practice, never happens once credentials are persisted).
+func (c *Client) WasJustProvisioned() bool {
+	v := c.justProvisioned
+	c.justProvisioned = false
+	return v
+}
+
 func (c *Client) metricsContext() context.Context {
 	c.metricsCtxMu.RLock()
 	defer c.metricsCtxMu.RUnlock()
@@ -481,6 +499,11 @@ func (c *Client) connectWithRetry() {
 func (c *Client) establishConnection() error {
 	ctx := context.Background()
 
+	// Exchange provisioning key for permanent credentials if needed.
+	if err := c.provisionIfNeeded(); err != nil {
+		return fmt.Errorf("failed to provision newt credentials: %w", err)
+	}
+
 	// Get token for authentication
 	token, err := c.getToken()
 	if err != nil {

websocket/config.go

@@ -1,16 +1,29 @@
 package websocket
 
 import (
+	"bytes"
+	"context"
+	"crypto/tls"
 	"encoding/json"
+	"fmt"
+	"io"
 	"log"
+	"net/http"
+	"net/url"
 	"os"
 	"path/filepath"
+	"regexp"
 	"runtime"
+	"strings"
+	"time"
 
 	"github.com/fosrl/newt/logger"
 )
 
-func getConfigPath(clientType string) string {
+func getConfigPath(clientType string, overridePath string) string {
+	if overridePath != "" {
+		return overridePath
+	}
 	configFile := os.Getenv("CONFIG_FILE")
 	if configFile == "" {
 		var configDir string
@@ -36,7 +49,7 @@ func getConfigPath(clientType string) string {
 
 func (c *Client) loadConfig() error {
 	originalConfig := *c.config // Store original config to detect changes
-	configPath := getConfigPath(c.clientType)
+	configPath := getConfigPath(c.clientType, c.configFilePath)
 
 	if c.config.ID != "" && c.config.Secret != "" && c.config.Endpoint != "" {
 		logger.Debug("Config already provided, skipping loading from file")
@@ -58,6 +71,11 @@ func (c *Client) loadConfig() error {
 		}
 		return err
 	}
+	if len(bytes.TrimSpace(data)) == 0 {
+		logger.Info("Config file at %s is empty, will initialize it with provided values", configPath)
+		c.configNeedsSave = true
+		return nil
+	}
 
 	var config Config
 	if err := json.Unmarshal(data, &config); err != nil {
@@ -83,6 +101,14 @@ func (c *Client) loadConfig() error {
 		c.config.Endpoint = config.Endpoint
 		c.baseURL = config.Endpoint
 	}
+	// Always load the provisioning key from the file if not already set
+	if c.config.ProvisioningKey == "" {
+		c.config.ProvisioningKey = config.ProvisioningKey
+	}
+	// Always load the name from the file if not already set
+	if c.config.Name == "" {
+		c.config.Name = config.Name
+	}
 
 	// Check if CLI args provided values that override file values
 	if (!fileHadID && originalConfig.ID != "") ||
@@ -105,7 +131,7 @@ func (c *Client) saveConfig() error {
 		return nil
 	}
 
-	configPath := getConfigPath(c.clientType)
+	configPath := getConfigPath(c.clientType, c.configFilePath)
 	data, err := json.MarshalIndent(c.config, "", "  ")
 	if err != nil {
 		return err
@@ -118,3 +144,139 @@ func (c *Client) saveConfig() error {
 	}
 	return err
 }
+
+// interpolateString replaces {{env.VAR}} tokens in s with the corresponding
+// environment variable values. Tokens that do not match a supported scheme are
+// left unchanged, mirroring the blueprint interpolation logic.
+func interpolateString(s string) string {
+	re := regexp.MustCompile(`\{\{([^}]+)\}\}`)
+	return re.ReplaceAllStringFunc(s, func(match string) string {
+		inner := strings.TrimSpace(match[2 : len(match)-2])
+		if strings.HasPrefix(inner, "env.") {
+			varName := strings.TrimPrefix(inner, "env.")
+			return os.Getenv(varName)
+		}
+		return match
+	})
+}
+
+// provisionIfNeeded checks whether a provisioning key is present and, if so,
+// exchanges it for a newt ID and secret by calling the registration endpoint.
+// On success the config is updated in-place and flagged for saving so that
+// subsequent runs use the permanent credentials directly.
+func (c *Client) provisionIfNeeded() error {
+	if c.config.ProvisioningKey == "" {
+		return nil
+	}
+
+	// If we already have both credentials there is nothing to provision.
+	if c.config.ID != "" && c.config.Secret != "" {
+		logger.Debug("Credentials already present, skipping provisioning")
+		return nil
+	}
+
+	logger.Info("Provisioning key found – exchanging for newt credentials...")
+
+	baseURL, err := url.Parse(c.baseURL)
+	if err != nil {
+		return fmt.Errorf("failed to parse base URL for provisioning: %w", err)
+	}
+	baseEndpoint := strings.TrimRight(baseURL.String(), "/")
+
+	// Interpolate any {{env.VAR}} tokens in the name before sending.
+	name := interpolateString(c.config.Name)
+
+	reqBody := map[string]interface{}{
+		"provisioningKey": c.config.ProvisioningKey,
+	}
+	if name != "" {
+		reqBody["name"] = name
+	}
+	jsonData, err := json.Marshal(reqBody)
+	if err != nil {
+		return fmt.Errorf("failed to marshal provisioning request: %w", err)
+	}
+
+	ctx, cancel := context.WithTimeout(context.Background(), 30*time.Second)
+	defer cancel()
+
+	req, err := http.NewRequestWithContext(
+		ctx,
+		"POST",
+		baseEndpoint+"/api/v1/auth/newt/register",
+		bytes.NewBuffer(jsonData),
+	)
+	if err != nil {
+		return fmt.Errorf("failed to create provisioning request: %w", err)
+	}
+	req.Header.Set("Content-Type", "application/json")
+	req.Header.Set("X-CSRF-Token", "x-csrf-protection")
+
+	// Mirror the TLS setup used by getToken so mTLS / self-signed CAs work.
+	var tlsCfg *tls.Config
+	if c.tlsConfig.ClientCertFile != "" || c.tlsConfig.ClientKeyFile != "" ||
+		len(c.tlsConfig.CAFiles) > 0 || c.tlsConfig.PKCS12File != "" {
+		tlsCfg, err = c.setupTLS()
+		if err != nil {
+			return fmt.Errorf("failed to setup TLS for provisioning: %w", err)
+		}
+	}
+	if os.Getenv("SKIP_TLS_VERIFY") == "true" {
+		if tlsCfg == nil {
+			tlsCfg = &tls.Config{}
+		}
+		tlsCfg.InsecureSkipVerify = true
+		logger.Debug("TLS certificate verification disabled for provisioning via SKIP_TLS_VERIFY")
+	}
+
+	httpClient := &http.Client{}
+	if tlsCfg != nil {
+		httpClient.Transport = &http.Transport{TLSClientConfig: tlsCfg}
+	}
+
+	resp, err := httpClient.Do(req)
+	if err != nil {
+		return fmt.Errorf("provisioning request failed: %w", err)
+	}
+	defer resp.Body.Close()
+
+	body, _ := io.ReadAll(resp.Body)
+	logger.Debug("Provisioning response body: %s", string(body))
+
+	if resp.StatusCode < 200 || resp.StatusCode >= 300 {
+		return fmt.Errorf("provisioning endpoint returned status %d: %s", resp.StatusCode, string(body))
+	}
+
+	var provResp ProvisioningResponse
+	if err := json.Unmarshal(body, &provResp); err != nil {
+		return fmt.Errorf("failed to decode provisioning response: %w", err)
+	}
+
+	if !provResp.Success {
+		return fmt.Errorf("provisioning failed: %s", provResp.Message)
+	}
+
+	if provResp.Data.NewtID == "" || provResp.Data.Secret == "" {
+		return fmt.Errorf("provisioning response is missing newt ID or secret")
+	}
+
+	logger.Info("Successfully provisioned – newt ID: %s", provResp.Data.NewtID)
+
+	// Persist the returned credentials and clear the one-time provisioning key
+	// so subsequent runs authenticate normally.
+	c.config.ID = provResp.Data.NewtID
+	c.config.Secret = provResp.Data.Secret
+	c.config.ProvisioningKey = ""
+	c.config.Name = ""
+	c.configNeedsSave = true
+	c.justProvisioned = true
+
+	// Save immediately so that if the subsequent connection attempt fails the
+	// provisioning key is already gone from disk and the next retry uses the
+	// permanent credentials instead of trying to provision again.
+	if err := c.saveConfig(); err != nil {
+		logger.Error("Failed to save config after provisioning: %v", err)
+	}
+
+	return nil
+}

websocket/config_test.go

@@ -0,0 +1,35 @@
+package websocket
+
+import (
+	"os"
+	"path/filepath"
+	"testing"
+)
+
+func TestLoadConfig_EmptyFileMarksConfigForSave(t *testing.T) {
+	t.Setenv("CONFIG_FILE", "")
+
+	tmpDir := t.TempDir()
+	configPath := filepath.Join(tmpDir, "config.json")
+	if err := os.WriteFile(configPath, []byte(""), 0o644); err != nil {
+		t.Fatalf("failed to create empty config file: %v", err)
+	}
+
+	client := &Client{
+		config: &Config{
+			Endpoint:        "https://example.com",
+			ProvisioningKey: "spk-test",
+		},
+		clientType:     "newt",
+		configFilePath: configPath,
+	}
+
+	if err := client.loadConfig(); err != nil {
+		t.Fatalf("loadConfig returned error for empty file: %v", err)
+	}
+
+	if !client.configNeedsSave {
+		t.Fatal("expected empty config file to mark configNeedsSave")
+	}
+}
+

websocket/types.go

@@ -1,10 +1,12 @@
 package websocket
 
 type Config struct {
-	ID            string `json:"id"`
-	Secret        string `json:"secret"`
-	Endpoint      string `json:"endpoint"`
-	TlsClientCert string `json:"tlsClientCert"`
+	ID              string `json:"id"`
+	Secret          string `json:"secret"`
+	Endpoint        string `json:"endpoint"`
+	TlsClientCert   string `json:"tlsClientCert"`
+	ProvisioningKey string `json:"provisioningKey,omitempty"`
+	Name            string `json:"name,omitempty"`
 }
 
 type TokenResponse struct {
@@ -16,8 +18,17 @@ type TokenResponse struct {
 	Message string `json:"message"`
 }
 
+type ProvisioningResponse struct {
+	Data struct {
+		NewtID string `json:"newtId"`
+		Secret string `json:"secret"`
+	} `json:"data"`
+	Success bool   `json:"success"`
+	Message string `json:"message"`
+}
+
 type WSMessage struct {
 	Type          string      `json:"type"`
 	Data          interface{} `json:"data"`
 	ConfigVersion int64       `json:"configVersion,omitempty"`
-}
+}