[Feature] Simplify Operator ID

CHANGELOG.md

@@ -8,6 +8,7 @@
 - (Maintenance) Bump Dependencies
 - (Feature) (Platform) EventsV1 Integration
 - (Feature) (Platform) Allows to opt out in the Inventory Telemetry 
+- (Feature) Simplify Operator ID Process
 
 ## [1.3.1](https://github.com/arangodb/kube-arangodb/tree/1.3.1) (2025-10-07)
 - (Documentation) Add ArangoPlatformStorage Docs & Examples

README.md

@@ -196,7 +196,6 @@ Flags:
       --http1.transport.keep-alive-timeout-short duration      Interval between keep-alive probes for an active network connection (default 100ms)
       --http1.transport.max-idle-conns int                     Maximum number of idle (keep-alive) connections across all hosts. Zero means no limit (default 100)
       --http1.transport.tls-handshake-timeout duration         Maximum amount of time to wait for a TLS handshake. Zero means no timeout (default 10s)
-      --image.discovery.status                                 Discover Operator Image from Pod Status by default. When disabled Pod Spec is used. (default true)
       --image.discovery.timeout duration                       Timeout for image discovery process (default 1m0s)
       --internal.scaling-integration                           Enable Scaling Integration
       --kubernetes.burst int                                   Burst for the k8s API (default 256)

cmd/cmd.go

@@ -162,7 +162,8 @@ var (
 		podSchedulingGracePeriod    time.Duration
 	}
 	operatorImageDiscovery struct {
-		timeout                time.Duration
+		timeout time.Duration
+		// Deprecated: Do not use this flag, as discovery method changed
 		defaultStatusDiscovery bool
 	}
 	operatorReconciliationRetry struct {
@@ -191,6 +192,12 @@ var (
 )
 
 func init() {
+	if err := initE(); err != nil {
+		panic(err.Error())
+	}
+}
+
+func initE() error {
 	var deprecatedStr string
 
 	f := cmdMain.Flags()
@@ -217,20 +224,14 @@ func init() {
 	f.BoolVar(&operatorOptions.enablePlatform, "operator.platform", false, "Enable to run the Platform operator")
 	f.BoolVar(&operatorOptions.enableScheduler, "operator.scheduler", false, "Enable to run the Scheduler operator")
 	f.BoolVar(&operatorOptions.enableK2KClusterSync, "operator.k2k-cluster-sync", false, "Enable to run the ListSimple operator")
-	f.MarkDeprecated("operator.k2k-cluster-sync", "Enabled within deployment operator")
 	f.BoolVar(&operatorOptions.versionOnly, "operator.version", false, "Enable only version endpoint in Operator")
 	f.StringVar(&deprecatedStr, "operator.alpine-image", "alpine:3.7", "Docker image used for alpine containers")
-	f.MarkDeprecated("operator.alpine-image", "Value is not used anymore")
 	f.StringVar(&deprecatedStr, "operator.metrics-exporter-image", "arangodb/arangodb-exporter:0.1.6", "Docker image used for metrics containers by default")
-	f.MarkDeprecated("operator.metrics-exporter-image", "Value is not used anymore")
 	f.StringVar(&deprecatedStr, "operator.arango-image", "arangodb/arangodb:latest", "Docker image used for arango by default")
-	f.MarkDeprecated("operator.arango-image", "Value is not used anymore")
 	f.BoolVar(&chaosOptions.allowed, "chaos.allowed", false, "Set to allow chaos in deployments. Only activated when allowed and enabled in deployment")
 	f.BoolVar(&operatorOptions.skipLeaderLabel, "leader.label.skip", false, "Skips Leader Label for the Pod")
 	f.BoolVar(&operatorOptions.singleMode, "mode.single", false, "Enable single mode in Operator. WARNING: There should be only one replica of Operator, otherwise Operator can take unexpected actions")
 	f.String("scope", "", "Define scope on which Operator works. Legacy - pre 1.1.0 scope with limited cluster access")
-	f.MarkDeprecated("scope", "Value is not used anymore")
-	f.MarkHidden("scope")
 	f.DurationVar(&operatorTimeouts.k8s, "timeout.k8s", globals.DefaultKubernetesTimeout, "The request timeout to the kubernetes")
 	f.DurationVar(&operatorTimeouts.arangoD, "timeout.arangod", globals.DefaultArangoDTimeout, "The request timeout to the ArangoDB")
 	f.DurationVar(&operatorTimeouts.arangoDCheck, "timeout.arangod-check", globals.DefaultArangoDCheckTimeout, "The version check request timeout to the ArangoDB")
@@ -257,24 +258,39 @@ func init() {
 	f.IntVar(&operatorBackup.concurrentUploads, "backup-concurrent-uploads", globals.DefaultBackupConcurrentUploads, "Number of concurrent uploads per deployment")
 	f.Uint64Var(&memoryLimit.hardLimit, "memory-limit", 0, "Define memory limit for hard shutdown and the dump of goroutines. Used for testing")
 	f.StringArrayVar(&metricsOptions.excludedMetricPrefixes, "metrics.excluded-prefixes", nil, "List of the excluded metrics prefixes")
-	f.BoolVar(&operatorImageDiscovery.defaultStatusDiscovery, "image.discovery.status", true, "Discover Operator Image from Pod Status by default. When disabled Pod Spec is used.")
+	f.BoolVar(&operatorImageDiscovery.defaultStatusDiscovery, "image.discovery.status", true, "Image discovery method is now determined by the deployment's ImageDiscoveryMode specification")
 	f.DurationVar(&operatorImageDiscovery.timeout, "image.discovery.timeout", time.Minute, "Timeout for image discovery process")
 	f.IntVar(&threads, "threads", 16, "Number of the worker threads")
-	if err := logging.Init(&cmdMain); err != nil {
-		panic(err.Error())
-	}
-	if err := features.Init(&cmdMain); err != nil {
-		panic(err.Error())
-	}
-	if err := agencyConfig.Init(&cmdMain); err != nil {
-		panic(err.Error())
+
+	if err := errors.Errors(
+		f.MarkDeprecated("operator.k2k-cluster-sync", "Enabled within deployment operator"),
+		f.MarkDeprecated("operator.alpine-image", "Value is not used anymore"),
+		f.MarkDeprecated("operator.metrics-exporter-image", "Value is not used anymore"),
+		f.MarkDeprecated("operator.arango-image", "Value is not used anymore"),
+		f.MarkDeprecated("scope", "Value is not used anymore"),
+		f.MarkDeprecated("image.discovery.status", "Value fetched from the Operator Spec"),
+	); err != nil {
+		return errors.Wrap(err, "Unable to mark flags as deprecated")
 	}
-	if err := reconcile.ActionsConfigGlobal.Init(&cmdMain); err != nil {
-		panic(err.Error())
+
+	if err := errors.Errors(
+		f.MarkHidden("scope"),
+		f.MarkHidden("image.discovery.status"),
+	); err != nil {
+		return errors.Wrap(err, "Unable to mark flags as hidden")
 	}
-	if err := operatorHTTP.InitConfiguration(&cmdMain); err != nil {
-		panic(err.Error())
+
+	if err := errors.Errors(
+		logging.Init(&cmdMain),
+		features.Init(&cmdMain),
+		agencyConfig.Init(&cmdMain),
+		reconcile.ActionsConfigGlobal.Init(&cmdMain),
+		operatorHTTP.InitConfiguration(&cmdMain),
+	); err != nil {
+		return errors.Wrap(err, "Unable to register secondary commands")
 	}
+
+	return nil
 }
 
 func Command() *cobra.Command {
@@ -590,7 +606,7 @@ func newOperatorConfigAndDeps(id, namespace, name string) (operator.Config, oper
 		Namespace:                   namespace,
 		PodName:                     name,
 		ServiceAccount:              serviceAccount,
-		OperatorImage:               image,
+		Image:                       image,
 		SkipLeaderLabel:             operatorOptions.skipLeaderLabel,
 		EnableDeployment:            operatorOptions.enableDeployment,
 		EnableDeploymentReplication: operatorOptions.enableDeploymentReplication,
@@ -633,18 +649,24 @@ func newOperatorConfigAndDeps(id, namespace, name string) (operator.Config, oper
 
 // getMyPodInfo looks up the image & service account of the pod with given name in given namespace
 // Returns image, serviceAccount, error.
-func getMyPodInfo(kubecli kubernetes.Interface, namespace, name string) (string, string, error) {
-	if image, sa, ok := getMyPodInfoWrap(kubecli, namespace, name, getMyImageInfoFunc(operatorImageDiscovery.defaultStatusDiscovery)); ok {
-		return image, sa, nil
-	}
+func getMyPodInfo(kubecli kubernetes.Interface, namespace, name string) (util.Image, string, error) {
+	var ret util.Image
+	var serviceAccount string
 
-	logger.Warn("Unable to discover image, fallback to second method")
+	if image, sa, ok := getMyPodInfoWrap(kubecli, namespace, name, getMyImageInfoFunc(false)); !ok {
+		return util.Image{}, "", errors.Errorf("Unable to discover Operator image from Spec")
+	} else {
+		ret.Image = image
+		serviceAccount = sa
+	}
 
-	if image, sa, ok := getMyPodInfoWrap(kubecli, namespace, name, getMyImageInfoFunc(!operatorImageDiscovery.defaultStatusDiscovery)); ok {
-		return image, sa, nil
+	if image, _, ok := getMyPodInfoWrap(kubecli, namespace, name, getMyImageInfoFunc(true)); ok {
+		ret.StatusImage = util.NewType(image)
+	} else {
+		logger.Warn("Unable to discover image from status")
 	}
 
-	return "", "", errors.Errorf("Unable to discover image")
+	return ret, serviceAccount, nil
 }
 
 func getMyPodInfoWrap(kubecli kubernetes.Interface, namespace, name string, imageFunc func(in *core.Pod) (string, bool)) (string, string, bool) {

cmd/cmd_pod_test.go

@@ -1,7 +1,7 @@
 //
 // DISCLAIMER
 //
-// Copyright 2024 ArangoDB GmbH, Cologne, Germany
+// Copyright 2025 ArangoDB GmbH, Cologne, Germany
 //
 // Licensed under the Apache License, Version 2.0 (the "License");
 // you may not use this file except in compliance with the License.
@@ -41,49 +41,52 @@ func Test_PodDiscovery(t *testing.T) {
 
 		Pod core.Pod
 
-		Image, ServiceAccount string
+		Image util.Image
 
-		Valid bool
+		ServiceAccount string
 
-		DefaultStatusDiscovery *bool
+		Valid bool
 	}
 
 	var testCases = []testCase{
+		//{
+		//	Name:  "Empty pod",
+		//	Valid: false,
+		//},
+		//{
+		//	Name:  "Not allowed containers",
+		//	Valid: false,
+		//	Pod: core.Pod{
+		//		ObjectMeta: meta.ObjectMeta{
+		//			Name:      "operator",
+		//			Namespace: tests.FakeNamespace,
+		//		},
+		//		Spec: core.PodSpec{
+		//			Containers: []core.Container{
+		//				{
+		//					Name:  "unknown",
+		//					Image: "image1",
+		//				},
+		//			},
+		//		},
+		//		Status: core.PodStatus{
+		//			ContainerStatuses: []core.ContainerStatus{
+		//				{
+		//					Name:    "unknown",
+		//					Image:   "image1",
+		//					ImageID: "image1",
+		//				},
+		//			},
+		//		},
+		//	},
+		//},
 		{
-			Name:  "Empty pod",
-			Valid: false,
-		},
-		{
-			Name:  "Not allowed containers",
-			Valid: false,
-			Pod: core.Pod{
-				ObjectMeta: meta.ObjectMeta{
-					Name:      "operator",
-					Namespace: tests.FakeNamespace,
-				},
-				Spec: core.PodSpec{
-					Containers: []core.Container{
-						{
-							Name:  "unknown",
-							Image: "image1",
-						},
-					},
-				},
-				Status: core.PodStatus{
-					ContainerStatuses: []core.ContainerStatus{
-						{
-							Name:    "unknown",
-							Image:   "image1",
-							ImageID: "image1",
-						},
-					},
-				},
+			Name:  "Allowed Status & Spec",
+			Valid: true,
+			Image: util.Image{
+				Image:       "image1",
+				StatusImage: util.NewType("image1"),
 			},
-		},
-		{
-			Name:           "Allowed Status & Spec",
-			Valid:          true,
-			Image:          "image1",
 			ServiceAccount: "sa",
 			Pod: core.Pod{
 				ObjectMeta: meta.ObjectMeta{
@@ -111,9 +114,12 @@ func Test_PodDiscovery(t *testing.T) {
 			},
 		},
 		{
-			Name:           "Allowed Status & Spec",
-			Valid:          true,
-			Image:          "imageStatusID1",
+			Name:  "Allowed Status & Spec",
+			Valid: true,
+			Image: util.Image{
+				Image:       "imageSpec1",
+				StatusImage: util.NewType("imageStatusID1"),
+			},
 			ServiceAccount: "sa",
 			Pod: core.Pod{
 				ObjectMeta: meta.ObjectMeta{
@@ -141,11 +147,13 @@ func Test_PodDiscovery(t *testing.T) {
 			},
 		},
 		{
-			Name:                   "Allowed Status & Spec - From Spec",
-			Valid:                  true,
-			Image:                  "imageSpec1",
-			ServiceAccount:         "sa",
-			DefaultStatusDiscovery: util.NewType(false),
+			Name:  "Allowed Status & Spec - From Spec",
+			Valid: true,
+			Image: util.Image{
+				Image:       "imageSpec1",
+				StatusImage: util.NewType("imageStatusID1"),
+			},
+			ServiceAccount: "sa",
 			Pod: core.Pod{
 				ObjectMeta: meta.ObjectMeta{
 					Name:      "operator",
@@ -172,9 +180,11 @@ func Test_PodDiscovery(t *testing.T) {
 			},
 		},
 		{
-			Name:           "Allowed Spec",
-			Valid:          true,
-			Image:          "imageSpec1",
+			Name:  "Allowed Spec",
+			Valid: true,
+			Image: util.Image{
+				Image: "imageSpec1",
+			},
 			ServiceAccount: "sa",
 			Pod: core.Pod{
 				ObjectMeta: meta.ObjectMeta{
@@ -193,9 +203,12 @@ func Test_PodDiscovery(t *testing.T) {
 			},
 		},
 		{
-			Name:           "Allowed Status & Spec - From Second Pod",
-			Valid:          true,
-			Image:          "imageStatusID2",
+			Name:  "Allowed Status & Spec - From Second Pod",
+			Valid: true,
+			Image: util.Image{
+				Image:       "imageSpec2",
+				StatusImage: util.NewType("imageStatusID2"),
+			},
 			ServiceAccount: "sa",
 			Pod: core.Pod{
 				ObjectMeta: meta.ObjectMeta{
@@ -232,11 +245,13 @@ func Test_PodDiscovery(t *testing.T) {
 			},
 		},
 		{
-			Name:                   "Allowed Status & Spec - From Second Pod Spec",
-			Valid:                  true,
-			Image:                  "imageSpec2",
-			ServiceAccount:         "sa",
-			DefaultStatusDiscovery: util.NewType(false),
+			Name:  "Allowed Status & Spec - From Second Pod Spec",
+			Valid: true,
+			Image: util.Image{
+				Image:       "imageSpec2",
+				StatusImage: util.NewType("imageStatusID2"),
+			},
+			ServiceAccount: "sa",
 			Pod: core.Pod{
 				ObjectMeta: meta.ObjectMeta{
 					Name:      "operator",
@@ -275,8 +290,6 @@ func Test_PodDiscovery(t *testing.T) {
 
 	for _, testCase := range testCases {
 		t.Run(testCase.Name, func(t *testing.T) {
-			operatorImageDiscovery.defaultStatusDiscovery = util.TypeOrDefault(testCase.DefaultStatusDiscovery, true)
-
 			c := kclient.NewFakeClientBuilder().Add(&testCase.Pod).Client()
 			image, sa, err := getMyPodInfo(c.Kubernetes(), tests.FakeNamespace, "operator")
 
@@ -286,7 +299,8 @@ func Test_PodDiscovery(t *testing.T) {
 				require.Empty(t, sa)
 			} else {
 				require.NoError(t, err)
-				require.Equal(t, testCase.Image, image)
+				require.Equal(t, testCase.Image.Image, image.Image)
+				tests.EqualPointers(t, testCase.Image.StatusImage, image.StatusImage)
 				require.Equal(t, testCase.ServiceAccount, sa)
 			}
 		})

docs/cli/arangodb_operator.md

@@ -78,7 +78,6 @@ Flags:
       --http1.transport.keep-alive-timeout-short duration      Interval between keep-alive probes for an active network connection (default 100ms)
       --http1.transport.max-idle-conns int                     Maximum number of idle (keep-alive) connections across all hosts. Zero means no limit (default 100)
       --http1.transport.tls-handshake-timeout duration         Maximum amount of time to wait for a TLS handshake. Zero means no timeout (default 10s)
-      --image.discovery.status                                 Discover Operator Image from Pod Status by default. When disabled Pod Spec is used. (default true)
       --image.discovery.timeout duration                       Timeout for image discovery process (default 1m0s)
       --internal.scaling-integration                           Enable Scaling Integration
       --kubernetes.burst int                                   Burst for the k8s API (default 256)

integrations/scheduler/v2/suite_test.go

@@ -32,6 +32,7 @@ import (
 	"github.com/arangodb/kube-arangodb/pkg/logging"
 	operator "github.com/arangodb/kube-arangodb/pkg/operatorV2"
 	"github.com/arangodb/kube-arangodb/pkg/operatorV2/event"
+	"github.com/arangodb/kube-arangodb/pkg/util"
 	"github.com/arangodb/kube-arangodb/pkg/util/k8sutil/helm"
 	"github.com/arangodb/kube-arangodb/pkg/util/kclient"
 	"github.com/arangodb/kube-arangodb/pkg/util/kclient/external"
@@ -102,7 +103,7 @@ func MockClient(t *testing.T, ctx context.Context, mods ...Mod) (pbSchedulerV2.S
 }
 
 func chartHandler(client kclient.Client, ns string) operator.Handler {
-	op := operator.NewOperator("mock", ns, "mock")
+	op := operator.NewOperator("mock", ns, util.Image{Image: "mock"})
 	recorder := event.NewEventRecorder("mock", client.Kubernetes())
 
 	return chart.Handler(op, recorder, client)

pkg/deployment/context_impl.go

@@ -84,7 +84,7 @@ func (d *Deployment) GetServerGroupIterator() reconciler.ServerGroupIterator {
 }
 
 func (d *Deployment) GetOperatorImage() string {
-	return d.config.OperatorImage
+	return d.config.Image.Get(d.GetSpec().ImageDiscoveryMode.Get() == api.DeploymentImageDiscoveryKubeletMode)
 }
 
 // GetNamespace returns the kubernetes namespace that contains

pkg/deployment/deployment.go

@@ -71,7 +71,7 @@ type Config struct {
 	ServiceAccount            string
 	AllowChaos                bool
 	ScalingIntegrationEnabled bool
-	OperatorImage             string
+	Image                     util.Image
 	ReconciliationDelay       time.Duration
 }