Merge pull request #14366 from rapphil/rapphil-fix-panic-ruler

Make rules Manager Update method  no-op after Close

rules/manager.go

@@ -190,10 +190,18 @@ func (m *Manager) Stop() {
 
 // Update the rule manager's state as the config requires. If
 // loading the new rules failed the old rule set is restored.
+// This method will no-op in case the manager is already stopped.
 func (m *Manager) Update(interval time.Duration, files []string, externalLabels labels.Labels, externalURL string, groupEvalIterationFunc GroupEvalIterationFunc) error {
 	m.mtx.Lock()
 	defer m.mtx.Unlock()
 
+	// We cannot update a stopped manager
+	select {
+	case <-m.done:
+		return nil
+	default:
+	}
+
 	groups, errs := m.LoadGroups(interval, externalLabels, externalURL, groupEvalIterationFunc, files...)
 
 	if errs != nil {

rules/manager_test.go

@@ -2099,6 +2099,23 @@ func TestBoundedRuleEvalConcurrency(t *testing.T) {
 	require.EqualValues(t, maxInflight.Load(), int32(maxConcurrency)+int32(groupCount))
 }
 
+func TestUpdateWhenStopped(t *testing.T) {
+	files := []string{"fixtures/rules.yaml"}
+	ruleManager := NewManager(&ManagerOptions{
+		Context: context.Background(),
+		Logger:  log.NewNopLogger(),
+	})
+	ruleManager.start()
+	err := ruleManager.Update(10*time.Second, files, labels.EmptyLabels(), "", nil)
+	require.NoError(t, err)
+	require.NotEmpty(t, ruleManager.groups)
+
+	ruleManager.Stop()
+	// Updates following a stop are no-op.
+	err = ruleManager.Update(10*time.Second, []string{}, labels.EmptyLabels(), "", nil)
+	require.NoError(t, err)
+}
+
 const artificialDelay = 250 * time.Millisecond
 
 func optsFactory(storage storage.Storage, maxInflight, inflightQueries *atomic.Int32, maxConcurrent int64) *ManagerOptions {