From bc02393c5a051ce3a065aed8cc365c1378a6572f Mon Sep 17 00:00:00 2001
From: =?UTF-8?q?Torjus=20H=C3=A5kestad?= <torjus@usit.uio.no>
Date: Sun, 8 Feb 2026 23:38:26 +0100
Subject: [PATCH] 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>
---
 cmd/homelab-deploy/main.go    |  2 +-
 internal/listener/listener.go | 11 +++++++++++
 internal/metrics/server.go    | 22 ++++++++++++++++++++--
 3 files changed, 32 insertions(+), 3 deletions(-)

diff --git a/cmd/homelab-deploy/main.go b/cmd/homelab-deploy/main.go
index 837014a..4c5bdde 100644
--- a/cmd/homelab-deploy/main.go
+++ b/cmd/homelab-deploy/main.go
@@ -16,7 +16,7 @@ import (
 	"github.com/urfave/cli/v3"
 )
 
-const version = "0.1.11"
+const version = "0.1.13"
 
 func main() {
 	app := &cli.Command{
diff --git a/internal/listener/listener.go b/internal/listener/listener.go
index b30357f..70b9ee3 100644
--- a/internal/listener/listener.go
+++ b/internal/listener/listener.go
@@ -270,6 +270,17 @@ func (l *Listener) handleDeployRequest(subject string, data []byte) {
 
 		// After a successful switch, signal restart so we pick up any new version
 		if req.Action == messages.ActionSwitch {
+			// Wait for metrics scrape before restarting (if metrics enabled)
+			if l.metricsServer != nil {
+				l.logger.Info("waiting for metrics scrape before restart")
+				select {
+				case <-l.metricsServer.ScrapeCh():
+					l.logger.Info("metrics scraped, proceeding with restart")
+				case <-time.After(60 * time.Second):
+					l.logger.Warn("no metrics scrape within timeout, proceeding with restart anyway")
+				}
+			}
+
 			select {
 			case l.restartCh <- struct{}{}:
 			default:
diff --git a/internal/metrics/server.go b/internal/metrics/server.go
index da6ebb8..886d463 100644
--- a/internal/metrics/server.go
+++ b/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)