feat: ability to set the status configMap name

cluster-autoscaler/FAQ.md

@@ -661,6 +661,7 @@ The following startup parameters are supported for cluster autoscaler:
 | `estimator` | Type of resource estimator to be used in scale up | binpacking
 | `expander` | Type of node group expander to be used in scale up.  | random
 | `write-status-configmap` | Should CA write status information to a configmap  | true
+| `status-config-map-name` | The name of the status ConfigMap that CA writes  | cluster-autoscaler-status
 | `max-inactivity` | Maximum time from last recorded autoscaler activity before automatic restart | 10 minutes
 | `max-failing-time` | Maximum time from last recorded successful autoscaler run before automatic restart | 15 minutes
 | `balance-similar-node-groups` | Detect similar node groups and balance the number of nodes between them | false

cluster-autoscaler/clusterstate/clusterstate_test.go

@@ -52,7 +52,7 @@ func TestOKWithScaleUp(t *testing.T) {
 	assert.NotNil(t, provider)
 
 	fakeClient := &fake.Clientset{}
-	fakeLogRecorder, _ := utils.NewStatusMapRecorder(fakeClient, "kube-system", kube_record.NewFakeRecorder(5), false)
+	fakeLogRecorder, _ := utils.NewStatusMapRecorder(fakeClient, "kube-system", kube_record.NewFakeRecorder(5), false, "my-cool-configmap")
 	clusterstate := NewClusterStateRegistry(provider, ClusterStateRegistryConfig{
 		MaxTotalUnreadyPercentage: 10,
 		OkTotalUnreadyCount:       1,
@@ -94,7 +94,7 @@ func TestEmptyOK(t *testing.T) {
 	assert.NotNil(t, provider)
 
 	fakeClient := &fake.Clientset{}
-	fakeLogRecorder, _ := utils.NewStatusMapRecorder(fakeClient, "kube-system", kube_record.NewFakeRecorder(5), false)
+	fakeLogRecorder, _ := utils.NewStatusMapRecorder(fakeClient, "kube-system", kube_record.NewFakeRecorder(5), false, "my-cool-configmap")
 	clusterstate := NewClusterStateRegistry(provider, ClusterStateRegistryConfig{
 		MaxTotalUnreadyPercentage: 10,
 		OkTotalUnreadyCount:       1,
@@ -135,7 +135,7 @@ func TestOKOneUnreadyNode(t *testing.T) {
 	assert.NotNil(t, provider)
 
 	fakeClient := &fake.Clientset{}
-	fakeLogRecorder, _ := utils.NewStatusMapRecorder(fakeClient, "kube-system", kube_record.NewFakeRecorder(5), false)
+	fakeLogRecorder, _ := utils.NewStatusMapRecorder(fakeClient, "kube-system", kube_record.NewFakeRecorder(5), false, "my-cool-configmap")
 	clusterstate := NewClusterStateRegistry(provider, ClusterStateRegistryConfig{
 		MaxTotalUnreadyPercentage: 10,
 		OkTotalUnreadyCount:       1,
@@ -173,7 +173,7 @@ func TestNodeWithoutNodeGroupDontCrash(t *testing.T) {
 	provider.AddNode("no_ng", noNgNode)
 
 	fakeClient := &fake.Clientset{}
-	fakeLogRecorder, _ := utils.NewStatusMapRecorder(fakeClient, "kube-system", kube_record.NewFakeRecorder(5), false)
+	fakeLogRecorder, _ := utils.NewStatusMapRecorder(fakeClient, "kube-system", kube_record.NewFakeRecorder(5), false, "my-cool-configmap")
 	clusterstate := NewClusterStateRegistry(provider, ClusterStateRegistryConfig{
 		MaxTotalUnreadyPercentage: 10,
 		OkTotalUnreadyCount:       1,
@@ -200,7 +200,7 @@ func TestOKOneUnreadyNodeWithScaleDownCandidate(t *testing.T) {
 	assert.NotNil(t, provider)
 
 	fakeClient := &fake.Clientset{}
-	fakeLogRecorder, _ := utils.NewStatusMapRecorder(fakeClient, "kube-system", kube_record.NewFakeRecorder(5), false)
+	fakeLogRecorder, _ := utils.NewStatusMapRecorder(fakeClient, "kube-system", kube_record.NewFakeRecorder(5), false, "my-cool-configmap")
 	clusterstate := NewClusterStateRegistry(provider, ClusterStateRegistryConfig{
 		MaxTotalUnreadyPercentage: 10,
 		OkTotalUnreadyCount:       1,
@@ -265,7 +265,7 @@ func TestMissingNodes(t *testing.T) {
 	assert.NotNil(t, provider)
 
 	fakeClient := &fake.Clientset{}
-	fakeLogRecorder, _ := utils.NewStatusMapRecorder(fakeClient, "kube-system", kube_record.NewFakeRecorder(5), false)
+	fakeLogRecorder, _ := utils.NewStatusMapRecorder(fakeClient, "kube-system", kube_record.NewFakeRecorder(5), false, "my-cool-configmap")
 	clusterstate := NewClusterStateRegistry(provider, ClusterStateRegistryConfig{
 		MaxTotalUnreadyPercentage: 10,
 		OkTotalUnreadyCount:       1,
@@ -307,7 +307,7 @@ func TestTooManyUnready(t *testing.T) {
 
 	assert.NotNil(t, provider)
 	fakeClient := &fake.Clientset{}
-	fakeLogRecorder, _ := utils.NewStatusMapRecorder(fakeClient, "kube-system", kube_record.NewFakeRecorder(5), false)
+	fakeLogRecorder, _ := utils.NewStatusMapRecorder(fakeClient, "kube-system", kube_record.NewFakeRecorder(5), false, "my-cool-configmap")
 	clusterstate := NewClusterStateRegistry(provider, ClusterStateRegistryConfig{
 		MaxTotalUnreadyPercentage: 10,
 		OkTotalUnreadyCount:       1,
@@ -331,7 +331,7 @@ func TestExpiredScaleUp(t *testing.T) {
 	assert.NotNil(t, provider)
 
 	fakeClient := &fake.Clientset{}
-	fakeLogRecorder, _ := utils.NewStatusMapRecorder(fakeClient, "kube-system", kube_record.NewFakeRecorder(5), false)
+	fakeLogRecorder, _ := utils.NewStatusMapRecorder(fakeClient, "kube-system", kube_record.NewFakeRecorder(5), false, "my-cool-configmap")
 	clusterstate := NewClusterStateRegistry(provider, ClusterStateRegistryConfig{
 		MaxTotalUnreadyPercentage: 10,
 		OkTotalUnreadyCount:       1,
@@ -357,7 +357,7 @@ func TestRegisterScaleDown(t *testing.T) {
 	assert.NotNil(t, provider)
 
 	fakeClient := &fake.Clientset{}
-	fakeLogRecorder, _ := utils.NewStatusMapRecorder(fakeClient, "kube-system", kube_record.NewFakeRecorder(5), false)
+	fakeLogRecorder, _ := utils.NewStatusMapRecorder(fakeClient, "kube-system", kube_record.NewFakeRecorder(5), false, "my-cool-configmap")
 	clusterstate := NewClusterStateRegistry(provider, ClusterStateRegistryConfig{
 		MaxTotalUnreadyPercentage: 10,
 		OkTotalUnreadyCount:       1,
@@ -410,7 +410,7 @@ func TestUpcomingNodes(t *testing.T) {
 
 	assert.NotNil(t, provider)
 	fakeClient := &fake.Clientset{}
-	fakeLogRecorder, _ := utils.NewStatusMapRecorder(fakeClient, "kube-system", kube_record.NewFakeRecorder(5), false)
+	fakeLogRecorder, _ := utils.NewStatusMapRecorder(fakeClient, "kube-system", kube_record.NewFakeRecorder(5), false, "my-cool-configmap")
 	clusterstate := NewClusterStateRegistry(provider, ClusterStateRegistryConfig{
 		MaxTotalUnreadyPercentage: 10,
 		OkTotalUnreadyCount:       1,
@@ -433,7 +433,7 @@ func TestIncorrectSize(t *testing.T) {
 	provider.AddNode("ng1", ng1_1)
 	assert.NotNil(t, provider)
 	fakeClient := &fake.Clientset{}
-	fakeLogRecorder, _ := utils.NewStatusMapRecorder(fakeClient, "kube-system", kube_record.NewFakeRecorder(5), false)
+	fakeLogRecorder, _ := utils.NewStatusMapRecorder(fakeClient, "kube-system", kube_record.NewFakeRecorder(5), false, "my-cool-configmap")
 	clusterstate := NewClusterStateRegistry(provider, ClusterStateRegistryConfig{
 		MaxTotalUnreadyPercentage: 10,
 		OkTotalUnreadyCount:       1,
@@ -469,7 +469,7 @@ func TestUnregisteredNodes(t *testing.T) {
 	provider.AddNode("ng1", ng1_2)
 
 	fakeClient := &fake.Clientset{}
-	fakeLogRecorder, _ := utils.NewStatusMapRecorder(fakeClient, "kube-system", kube_record.NewFakeRecorder(5), false)
+	fakeLogRecorder, _ := utils.NewStatusMapRecorder(fakeClient, "kube-system", kube_record.NewFakeRecorder(5), false, "my-cool-configmap")
 	clusterstate := NewClusterStateRegistry(provider, ClusterStateRegistryConfig{
 		MaxTotalUnreadyPercentage: 10,
 		OkTotalUnreadyCount:       1,
@@ -621,7 +621,7 @@ func TestScaleUpBackoff(t *testing.T) {
 	assert.NotNil(t, provider)
 
 	fakeClient := &fake.Clientset{}
-	fakeLogRecorder, _ := utils.NewStatusMapRecorder(fakeClient, "kube-system", kube_record.NewFakeRecorder(5), false)
+	fakeLogRecorder, _ := utils.NewStatusMapRecorder(fakeClient, "kube-system", kube_record.NewFakeRecorder(5), false, "my-cool-configmap")
 	clusterstate := NewClusterStateRegistry(provider, ClusterStateRegistryConfig{
 		MaxTotalUnreadyPercentage: 10,
 		OkTotalUnreadyCount:       1,
@@ -689,7 +689,7 @@ func TestGetClusterSize(t *testing.T) {
 	provider.AddNode("notAutoscaledNode", notAutoscaledNode)
 
 	fakeClient := &fake.Clientset{}
-	fakeLogRecorder, _ := utils.NewStatusMapRecorder(fakeClient, "kube-system", kube_record.NewFakeRecorder(5), false)
+	fakeLogRecorder, _ := utils.NewStatusMapRecorder(fakeClient, "kube-system", kube_record.NewFakeRecorder(5), false, "my-cool-configmap")
 	clusterstate := NewClusterStateRegistry(provider, ClusterStateRegistryConfig{
 		MaxTotalUnreadyPercentage: 10,
 		OkTotalUnreadyCount:       1,
@@ -731,7 +731,7 @@ func TestUpdateScaleUp(t *testing.T) {
 	provider := testprovider.NewTestCloudProvider(nil, nil)
 	provider.AddNodeGroup("ng1", 1, 10, 5)
 	fakeClient := &fake.Clientset{}
-	fakeLogRecorder, _ := utils.NewStatusMapRecorder(fakeClient, "kube-system", kube_record.NewFakeRecorder(5), false)
+	fakeLogRecorder, _ := utils.NewStatusMapRecorder(fakeClient, "kube-system", kube_record.NewFakeRecorder(5), false, "my-cool-configmap")
 	clusterstate := NewClusterStateRegistry(
 		provider,
 		ClusterStateRegistryConfig{
@@ -809,7 +809,7 @@ func TestScaleUpFailures(t *testing.T) {
 	assert.NotNil(t, provider)
 
 	fakeClient := &fake.Clientset{}
-	fakeLogRecorder, _ := utils.NewStatusMapRecorder(fakeClient, "kube-system", kube_record.NewFakeRecorder(5), false)
+	fakeLogRecorder, _ := utils.NewStatusMapRecorder(fakeClient, "kube-system", kube_record.NewFakeRecorder(5), false, "my-cool-configmap")
 	clusterstate := NewClusterStateRegistry(provider, ClusterStateRegistryConfig{}, fakeLogRecorder, newBackoff())
 
 	clusterstate.RegisterFailedScaleUp(provider.GetNodeGroup("ng1"), metrics.Timeout, now)

cluster-autoscaler/clusterstate/utils/status.go

@@ -32,8 +32,6 @@ import (
 )
 
 const (
-	// StatusConfigMapName is the name of ConfigMap with status.
-	StatusConfigMapName = "cluster-autoscaler-status"
 	// ConfigMapLastUpdatedKey is the name of annotation informing about status ConfigMap last update.
 	ConfigMapLastUpdatedKey = "cluster-autoscaler.kubernetes.io/last-updated"
 	// ConfigMapLastUpdateFormat it the timestamp format used for last update annotation in status ConfigMap
@@ -64,11 +62,11 @@ func (ler *LogEventRecorder) Eventf(eventtype, reason, message string, args ...i
 // NewStatusMapRecorder creates a LogEventRecorder creating events on status configmap.
 // If the configmap doesn't exist it will be created (with 'Initializing' status).
 // If active == false the map will not be created and no events will be recorded.
-func NewStatusMapRecorder(kubeClient kube_client.Interface, namespace string, recorder record.EventRecorder, active bool) (*LogEventRecorder, error) {
+func NewStatusMapRecorder(kubeClient kube_client.Interface, namespace string, recorder record.EventRecorder, active bool, statusConfigMapName string) (*LogEventRecorder, error) {
 	var mapObj runtime.Object
 	var err error
 	if active {
-		mapObj, err = WriteStatusConfigMap(kubeClient, namespace, "Initializing", nil)
+		mapObj, err = WriteStatusConfigMap(kubeClient, namespace, "Initializing", nil, statusConfigMapName)
 		if err != nil {
 			return nil, errors.New("Failed to init status ConfigMap")
 		}
@@ -83,14 +81,14 @@ func NewStatusMapRecorder(kubeClient kube_client.Interface, namespace string, re
 // WriteStatusConfigMap writes updates status ConfigMap with a given message or creates a new
 // ConfigMap if it doesn't exist. If logRecorder is passed and configmap update is successful
 // logRecorder's internal reference will be updated.
-func WriteStatusConfigMap(kubeClient kube_client.Interface, namespace string, msg string, logRecorder *LogEventRecorder) (*apiv1.ConfigMap, error) {
+func WriteStatusConfigMap(kubeClient kube_client.Interface, namespace string, msg string, logRecorder *LogEventRecorder, statusConfigMapName string) (*apiv1.ConfigMap, error) {
 	statusUpdateTime := time.Now().Format(ConfigMapLastUpdateFormat)
 	statusMsg := fmt.Sprintf("Cluster-autoscaler status at %s:\n%v", statusUpdateTime, msg)
 	var configMap *apiv1.ConfigMap
 	var getStatusError, writeStatusError error
 	var errMsg string
 	maps := kubeClient.CoreV1().ConfigMaps(namespace)
-	configMap, getStatusError = maps.Get(StatusConfigMapName, metav1.GetOptions{})
+	configMap, getStatusError = maps.Get(statusConfigMapName, metav1.GetOptions{})
 	if getStatusError == nil {
 		configMap.Data["status"] = statusMsg
 		if configMap.ObjectMeta.Annotations == nil {
@@ -102,7 +100,7 @@ func WriteStatusConfigMap(kubeClient kube_client.Interface, namespace string, ms
 		configMap = &apiv1.ConfigMap{
 			ObjectMeta: metav1.ObjectMeta{
 				Namespace: namespace,
-				Name:      StatusConfigMapName,
+				Name:      statusConfigMapName,
 				Annotations: map[string]string{
 					ConfigMapLastUpdatedKey: statusUpdateTime,
 				},
@@ -132,9 +130,9 @@ func WriteStatusConfigMap(kubeClient kube_client.Interface, namespace string, ms
 }
 
 // DeleteStatusConfigMap deletes status configmap
-func DeleteStatusConfigMap(kubeClient kube_client.Interface, namespace string) error {
+func DeleteStatusConfigMap(kubeClient kube_client.Interface, namespace string, statusConfigMapName string) error {
 	maps := kubeClient.CoreV1().ConfigMaps(namespace)
-	err := maps.Delete(StatusConfigMapName, &metav1.DeleteOptions{})
+	err := maps.Delete(statusConfigMapName, &metav1.DeleteOptions{})
 	if err != nil {
 		klog.Error("Failed to delete status configmap")
 	}

cluster-autoscaler/clusterstate/utils/status_test.go

@@ -48,7 +48,7 @@ func setUpTest(t *testing.T) *testInfo {
 		configMap: &apiv1.ConfigMap{
 			ObjectMeta: metav1.ObjectMeta{
 				Namespace: namespace,
-				Name:      StatusConfigMapName,
+				Name:      "my-cool-configmap",
 			},
 			Data: map[string]string{},
 		},
@@ -61,7 +61,7 @@ func setUpTest(t *testing.T) *testInfo {
 	result.client.Fake.AddReactor("get", "configmaps", func(action core.Action) (bool, runtime.Object, error) {
 		get := action.(core.GetAction)
 		assert.Equal(result.t, namespace, get.GetNamespace())
-		assert.Equal(result.t, StatusConfigMapName, get.GetName())
+		assert.Equal(result.t, "my-cool-configmap", get.GetName())
 		result.getCalled = true
 		if result.getError != nil {
 			return true, nil, result.getError
@@ -78,7 +78,7 @@ func setUpTest(t *testing.T) *testInfo {
 		create := action.(core.CreateAction)
 		assert.Equal(result.t, namespace, create.GetNamespace())
 		configMap := create.GetObject().(*apiv1.ConfigMap)
-		assert.Equal(result.t, StatusConfigMapName, configMap.ObjectMeta.Name)
+		assert.Equal(result.t, "my-cool-configmap", configMap.ObjectMeta.Name)
 		result.createCalled = true
 		return true, configMap, nil
 	})

cluster-autoscaler/config/autoscaling_options.go

@@ -109,6 +109,8 @@ type AutoscalingOptions struct {
 	NodeDeletionDelayTimeout time.Duration
 	// WriteStatusConfigMap tells if the status information should be written to a ConfigMap
 	WriteStatusConfigMap bool
+	// StaticConfigMapName
+	StatusConfigMapName string
 	// BalanceSimilarNodeGroups enables logic that identifies node groups with similar machines and tries to balance node count between them.
 	BalanceSimilarNodeGroups bool
 	// ConfigNamespace is the namespace cluster-autoscaler is running in and all related configmaps live in

cluster-autoscaler/context/autoscaling_context.go

@@ -103,12 +103,12 @@ func NewAutoscalingKubeClients(opts config.AutoscalingOptions, kubeClient, event
 	listerRegistryStopChannel := make(chan struct{})
 	listerRegistry := kube_util.NewListerRegistryWithDefaultListers(kubeClient, listerRegistryStopChannel)
 	kubeEventRecorder := kube_util.CreateEventRecorder(eventsKubeClient)
-	logRecorder, err := utils.NewStatusMapRecorder(kubeClient, opts.ConfigNamespace, kubeEventRecorder, opts.WriteStatusConfigMap)
+	logRecorder, err := utils.NewStatusMapRecorder(kubeClient, opts.ConfigNamespace, kubeEventRecorder, opts.WriteStatusConfigMap, opts.StatusConfigMapName)
 	if err != nil {
 		klog.Error("Failed to initialize status configmap, unable to write status events")
 		// Get a dummy, so we can at least safely call the methods
 		// TODO(maciekpytel): recover from this after successful status configmap update?
-		logRecorder, _ = utils.NewStatusMapRecorder(eventsKubeClient, opts.ConfigNamespace, kubeEventRecorder, false)
+		logRecorder, _ = utils.NewStatusMapRecorder(eventsKubeClient, opts.ConfigNamespace, kubeEventRecorder, false, opts.StatusConfigMapName)
 	}
 
 	return &AutoscalingKubeClients{

cluster-autoscaler/core/scale_test_common.go

@@ -111,7 +111,7 @@ func NewScaleTestAutoscalingContext(
 	listers kube_util.ListerRegistry, provider cloudprovider.CloudProvider,
 	processorCallbacks processor_callbacks.ProcessorCallbacks) context.AutoscalingContext {
 	fakeRecorder := kube_record.NewFakeRecorder(5)
-	fakeLogRecorder, _ := utils.NewStatusMapRecorder(fakeClient, "kube-system", fakeRecorder, false)
+	fakeLogRecorder, _ := utils.NewStatusMapRecorder(fakeClient, "kube-system", fakeRecorder, false, "my-cool-configmap")
 	// Ignoring error here is safe - if a test doesn't specify valid estimatorName,
 	// it either doesn't need one, or should fail when it turns out to be nil.
 	estimatorBuilder, _ := estimator.NewEstimatorBuilder(options.EstimatorName)

cluster-autoscaler/core/static_autoscaler.go

@@ -244,7 +244,7 @@ func (a *StaticAutoscaler) RunOnce(currentTime time.Time) errors.AutoscalerError
 		if autoscalingContext.WriteStatusConfigMap {
 			status := a.clusterStateRegistry.GetStatus(currentTime)
 			utils.WriteStatusConfigMap(autoscalingContext.ClientSet, autoscalingContext.ConfigNamespace,
-				status.GetReadableString(), a.AutoscalingContext.LogRecorder)
+				status.GetReadableString(), a.AutoscalingContext.LogRecorder, a.AutoscalingContext.StatusConfigMapName)
 		}
 
 		// This deferred processor execution allows the processors to handle a situation when a scale-(up|down)
@@ -631,7 +631,7 @@ func (a *StaticAutoscaler) ExitCleanUp() {
 	if !a.AutoscalingContext.WriteStatusConfigMap {
 		return
 	}
-	utils.DeleteStatusConfigMap(a.AutoscalingContext.ClientSet, a.AutoscalingContext.ConfigNamespace)
+	utils.DeleteStatusConfigMap(a.AutoscalingContext.ClientSet, a.AutoscalingContext.ConfigNamespace, a.AutoscalingContext.StatusConfigMapName)
 
 	a.clusterStateRegistry.Stop()
 }
@@ -695,7 +695,7 @@ func (a *StaticAutoscaler) onEmptyCluster(status string, emitEvent bool) {
 	metrics.UpdateClusterSafeToAutoscale(false)
 	metrics.UpdateNodesCount(0, 0, 0, 0, 0)
 	if a.AutoscalingContext.WriteStatusConfigMap {
-		utils.WriteStatusConfigMap(a.AutoscalingContext.ClientSet, a.AutoscalingContext.ConfigNamespace, status, a.AutoscalingContext.LogRecorder)
+		utils.WriteStatusConfigMap(a.AutoscalingContext.ClientSet, a.AutoscalingContext.ConfigNamespace, status, a.AutoscalingContext.LogRecorder, a.AutoscalingContext.StatusConfigMapName)
 	}
 	if emitEvent {
 		a.AutoscalingContext.LogRecorder.Eventf(apiv1.EventTypeWarning, "ClusterUnhealthy", status)

cluster-autoscaler/core/static_autoscaler_test.go

@@ -1354,7 +1354,7 @@ func TestRemoveFixNodeTargetSize(t *testing.T) {
 	provider.AddNode("ng1", ng1_1)
 
 	fakeClient := &fake.Clientset{}
-	fakeLogRecorder, _ := clusterstate_utils.NewStatusMapRecorder(fakeClient, "kube-system", kube_record.NewFakeRecorder(5), false)
+	fakeLogRecorder, _ := clusterstate_utils.NewStatusMapRecorder(fakeClient, "kube-system", kube_record.NewFakeRecorder(5), false, "my-cool-configmap")
 	clusterState := clusterstate.NewClusterStateRegistry(provider, clusterstate.ClusterStateRegistryConfig{
 		MaxTotalUnreadyPercentage: 10,
 		OkTotalUnreadyCount:       1,
@@ -1400,7 +1400,7 @@ func TestRemoveOldUnregisteredNodes(t *testing.T) {
 	provider.AddNode("ng1", ng1_2)
 
 	fakeClient := &fake.Clientset{}
-	fakeLogRecorder, _ := clusterstate_utils.NewStatusMapRecorder(fakeClient, "kube-system", kube_record.NewFakeRecorder(5), false)
+	fakeLogRecorder, _ := clusterstate_utils.NewStatusMapRecorder(fakeClient, "kube-system", kube_record.NewFakeRecorder(5), false, "my-cool-configmap")
 	clusterState := clusterstate.NewClusterStateRegistry(provider, clusterstate.ClusterStateRegistryConfig{
 		MaxTotalUnreadyPercentage: 10,
 		OkTotalUnreadyCount:       1,