From 357e6520449004565aeb4a7ed6c05f74c5cd1d87 Mon Sep 17 00:00:00 2001
From: Lukasz Mierzwa <l.mierzwa@gmail.com>
Date: Wed, 2 Jul 2025 16:01:42 +0100
Subject: [PATCH] Add a test for a rare shutdown hang

When doing a config reload that need to stop some providers while also sending SIGTERM to Prometheus at the same time can sometimes hang

1: sync.WaitGroup.Wait [83 minutes] [Created by run.(*Group).Run in goroutine 1 @ group.go:37]
    sync         sema.go:110              runtime_SemacquireWaitGroup(*uint32(#166))
    sync         waitgroup.go:118         (*WaitGroup).Wait(*WaitGroup(#23))
    discovery    manager.go:276           (*Manager).ApplyConfig(#23, #167)
    main         main.go:964              main.func5(#120)
    main         main.go:1505             reloadConfig({#183, 0x1b}, 1, #40, #43, #50, {#31, 0xa, 0})
    main         main.go:1182             main.func22()
    run          group.go:38              (*Group).Run.func1(*Group(#26), #51)

Add a test for it.

Signed-off-by: Lukasz Mierzwa <l.mierzwa@gmail.com>
---
 discovery/manager_test.go | 50 +++++++++++++++++++++++++++++++++++++++
 1 file changed, 50 insertions(+)

diff --git a/discovery/manager_test.go b/discovery/manager_test.go
index 1dd10baf47..38a93be9f4 100644
--- a/discovery/manager_test.go
+++ b/discovery/manager_test.go
@@ -1562,3 +1562,53 @@ func TestUnregisterMetrics(t *testing.T) {
 		cancel()
 	}
 }
+
+// Calling ApplyConfig() that removes providers at the same time as shutting down
+// the manager should not hang.
+func TestConfigReloadAndShutdownRace(t *testing.T) {
+	reg := prometheus.NewRegistry()
+	_, sdMetrics := NewTestMetrics(t, reg)
+
+	mgrCtx, mgrCancel := context.WithCancel(context.Background())
+	discoveryManager := NewManager(mgrCtx, promslog.NewNopLogger(), reg, sdMetrics)
+	require.NotNil(t, discoveryManager)
+	discoveryManager.updatert = 100 * time.Millisecond
+
+	var wgDiscovery sync.WaitGroup
+	wgDiscovery.Add(1)
+	go func() {
+		discoveryManager.Run()
+		wgDiscovery.Done()
+	}()
+	time.Sleep(time.Millisecond * 200)
+
+	var wgBg sync.WaitGroup
+	updateChan := discoveryManager.SyncCh()
+	wgBg.Add(1)
+	ctx, cancel := context.WithCancel(context.Background())
+	go func() {
+		defer wgBg.Done()
+		select {
+		case <-ctx.Done():
+			return
+		case <-updateChan:
+		}
+	}()
+
+	c := map[string]Configs{
+		"prometheus": {staticConfig("bar:9090")},
+	}
+	discoveryManager.ApplyConfig(c)
+
+	delete(c, "prometheus")
+	wgBg.Add(1)
+	go func() {
+		discoveryManager.ApplyConfig(c)
+		wgBg.Done()
+	}()
+	mgrCancel()
+	wgDiscovery.Wait()
+
+	cancel()
+	wgBg.Wait()
+}