fix: wait for metrics scrape before restarting after switch deployment

After a successful switch deployment, the listener now waits for Prometheus
to scrape the /metrics endpoint before exiting for restart. This ensures
deployment metrics are captured before the process restarts and resets
in-memory counters. Falls back to a 60 second timeout if no scrape occurs.

Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>

internal/metrics/server.go

@@ -23,6 +23,7 @@ type Server struct {
 	registry   *prometheus.Registry
 	collector  *Collector
 	logger     *slog.Logger
+	scrapeCh   chan struct{}
 }
 
 // NewServer creates a new metrics server.
@@ -35,9 +36,20 @@ func NewServer(cfg ServerConfig) *Server {
 	registry := prometheus.NewRegistry()
 	collector := NewCollector(registry)
 
-	mux := http.NewServeMux()
-	mux.Handle("/metrics", promhttp.HandlerFor(registry, promhttp.HandlerOpts{
+	scrapeCh := make(chan struct{})
+
+	metricsHandler := promhttp.HandlerFor(registry, promhttp.HandlerOpts{
 		Registry: registry,
+	})
+
+	mux := http.NewServeMux()
+	mux.Handle("/metrics", http.HandlerFunc(func(w http.ResponseWriter, r *http.Request) {
+		metricsHandler.ServeHTTP(w, r)
+		// Signal that a scrape occurred (non-blocking)
+		select {
+		case scrapeCh <- struct{}{}:
+		default:
+		}
 	}))
 	mux.HandleFunc("/health", func(w http.ResponseWriter, _ *http.Request) {
 		w.WriteHeader(http.StatusOK)
@@ -53,6 +65,7 @@ func NewServer(cfg ServerConfig) *Server {
 		registry:  registry,
 		collector: collector,
 		logger:    logger,
+		scrapeCh:  scrapeCh,
 	}
 }
 
@@ -61,6 +74,11 @@ func (s *Server) Collector() *Collector {
 	return s.collector
 }
 
+// ScrapeCh returns a channel that receives a signal each time the metrics endpoint is scraped.
+func (s *Server) ScrapeCh() <-chan struct{} {
+	return s.scrapeCh
+}
+
 // Start starts the HTTP server in a goroutine.
 func (s *Server) Start() error {
 	s.logger.Info("starting metrics server", "addr", s.httpServer.Addr)