feat: add DGDR Status .state enum (#6396)

deploy/helm/charts/crds/templates/nvidia.com_dynamographdeploymentrequests.yaml

@@ -57,7 +57,7 @@ spec:
             specific performance and resource constraints, enabling SLA-driven deployments.
 
             Lifecycle:
-             1. Initial → Pending: Validates spec and prepares for profiling
+             1. Initializing → Pending: Validates spec and prepares for profiling
              2. Pending → Profiling: Creates and runs profiling job (online or AIC)
              3. Profiling → Ready/Deploying: Generates DGD spec after profiling completes
              4. Deploying → Ready: When autoApply=true, monitors DGD until Ready
@@ -451,11 +451,19 @@ spec:
                     Format: "configmap/\<name\>"
                   type: string
                 state:
-                  description: |-
-                    State is a high-level textual status of the deployment request lifecycle.
-                    Possible values: "", "Pending", "Profiling", "Deploying", "Ready", "DeploymentDeleted", "Failed"
-                    Empty string ("") represents the initial state before initialization.
+                  default: Initializing
+                  description: State is a high-level textual status of the deployment request lifecycle.
+                  enum:
+                    - Initializing
+                    - Pending
+                    - Profiling
+                    - Deploying
+                    - Ready
+                    - DeploymentDeleted
+                    - Failed
                   type: string
+              required:
+                - state
               type: object
           type: object
       served: true

deploy/operator/api/v1alpha1/dynamographdeploymentrequest_types.go

@@ -28,8 +28,6 @@ import (
 	apiextensionsv1 "k8s.io/apiextensions-apiserver/pkg/apis/apiextensions/v1"
 	metav1 "k8s.io/apimachinery/pkg/apis/meta/v1"
 	runtime "k8s.io/apimachinery/pkg/runtime"
-
-	"github.com/ai-dynamo/dynamo/deploy/operator/internal/consts"
 )
 
 // EDIT THIS FILE!  THIS IS SCAFFOLDING FOR YOU TO OWN!
@@ -96,6 +94,19 @@ type ProfilingConfigSpec struct {
 	NodeSelector map[string]string `json:"nodeSelector,omitempty"`
 }
 
+// +kubebuilder:validation:Enum=Initializing;Pending;Profiling;Deploying;Ready;DeploymentDeleted;Failed
+type DGDRState string
+
+const (
+	DGDRStateInitializing      DGDRState = "Initializing"
+	DGDRStatePending           DGDRState = "Pending"
+	DGDRStateProfiling         DGDRState = "Profiling"
+	DGDRStateDeploying         DGDRState = "Deploying"
+	DGDRStateReady             DGDRState = "Ready"
+	DGDRStateDeploymentDeleted DGDRState = "DeploymentDeleted"
+	DGDRStateFailed            DGDRState = "Failed"
+)
+
 // DeploymentOverridesSpec allows users to customize metadata for auto-created DynamoGraphDeployments.
 // When autoApply is enabled, these overrides are applied to the generated DGD resource.
 type DeploymentOverridesSpec struct {
@@ -209,9 +220,8 @@ type DeploymentStatus struct {
 // The controller updates this status as the DGDR progresses through its lifecycle.
 type DynamoGraphDeploymentRequestStatus struct {
 	// State is a high-level textual status of the deployment request lifecycle.
-	// Possible values: "", "Pending", "Profiling", "Deploying", "Ready", "DeploymentDeleted", "Failed"
-	// Empty string ("") represents the initial state before initialization.
-	State string `json:"state,omitempty"`
+	// +kubebuilder:default=Initializing
+	State DGDRState `json:"state"`
 
 	// Backend is extracted from profilingConfig.config.engine.backend for display purposes.
 	// This field is populated by the controller and shown in kubectl output.
@@ -253,7 +263,7 @@ type DynamoGraphDeploymentRequestStatus struct {
 // specific performance and resource constraints, enabling SLA-driven deployments.
 //
 // Lifecycle:
-//  1. Initial → Pending: Validates spec and prepares for profiling
+//  1. Initializing → Pending: Validates spec and prepares for profiling
 //  2. Pending → Profiling: Creates and runs profiling job (online or AIC)
 //  3. Profiling → Ready/Deploying: Generates DGD spec after profiling completes
 //  4. Deploying → Ready: When autoApply=true, monitors DGD until Ready
@@ -283,16 +293,13 @@ type DynamoGraphDeploymentRequest struct {
 }
 
 // SetState updates the State field in the DGDR status.
-func (s *DynamoGraphDeploymentRequest) SetState(state string) {
+func (s *DynamoGraphDeploymentRequest) SetState(state DGDRState) {
 	s.Status.State = state
 }
 
 // GetState returns the current lifecycle state
 func (d *DynamoGraphDeploymentRequest) GetState() string {
-	if d.Status.State == "" {
-		return consts.ResourceStateUnknown
-	}
-	return d.Status.State
+	return string(d.Status.State)
 }
 
 // GetSpec returns the spec of this DGDR as a generic interface.

deploy/operator/config/crd/bases/nvidia.com_dynamographdeploymentrequests.yaml

@@ -57,7 +57,7 @@ spec:
             specific performance and resource constraints, enabling SLA-driven deployments.
 
             Lifecycle:
-             1. Initial → Pending: Validates spec and prepares for profiling
+             1. Initializing → Pending: Validates spec and prepares for profiling
              2. Pending → Profiling: Creates and runs profiling job (online or AIC)
              3. Profiling → Ready/Deploying: Generates DGD spec after profiling completes
              4. Deploying → Ready: When autoApply=true, monitors DGD until Ready
@@ -451,11 +451,19 @@ spec:
                     Format: "configmap/\<name\>"
                   type: string
                 state:
-                  description: |-
-                    State is a high-level textual status of the deployment request lifecycle.
-                    Possible values: "", "Pending", "Profiling", "Deploying", "Ready", "DeploymentDeleted", "Failed"
-                    Empty string ("") represents the initial state before initialization.
+                  default: Initializing
+                  description: State is a high-level textual status of the deployment request lifecycle.
+                  enum:
+                    - Initializing
+                    - Pending
+                    - Profiling
+                    - Deploying
+                    - Ready
+                    - DeploymentDeleted
+                    - Failed
                   type: string
+              required:
+                - state
               type: object
           type: object
       served: true

deploy/operator/internal/controller/dynamographdeploymentrequest_controller.go

@@ -53,15 +53,6 @@ import (
 )
 
 const (
-	// DGDR state constants
-	DGDRStateEmpty             = ""
-	DGDRStatePending           = "Pending"
-	DGDRStateProfiling         = "Profiling"
-	DGDRStateDeploying         = "Deploying"
-	DGDRStateReady             = "Ready"
-	DGDRStateDeploymentDeleted = "DeploymentDeleted"
-	DGDRStateFailed            = "Failed"
-
 	// Condition types
 	ConditionTypeValidation      = "Validation"
 	ConditionTypeProfiling       = "Profiling"
@@ -373,8 +364,8 @@ func (r *DynamoGraphDeploymentRequestReconciler) Reconcile(ctx context.Context,
 	// Check for spec changes (immutability enforcement)
 	if dgdr.Status.ObservedGeneration > 0 && dgdr.Status.ObservedGeneration != dgdr.Generation {
 		// Spec changed after initial processing
-		if dgdr.Status.State == DGDRStateProfiling || dgdr.Status.State == DGDRStateDeploying ||
-			dgdr.Status.State == DGDRStateReady || dgdr.Status.State == DGDRStateDeploymentDeleted {
+		if dgdr.Status.State == nvidiacomv1alpha1.DGDRStateProfiling || dgdr.Status.State == nvidiacomv1alpha1.DGDRStateDeploying ||
+			dgdr.Status.State == nvidiacomv1alpha1.DGDRStateReady || dgdr.Status.State == nvidiacomv1alpha1.DGDRStateDeploymentDeleted {
 			logger.Info("Spec change detected in immutable state",
 				"state", dgdr.Status.State,
 				"observedGeneration", dgdr.Status.ObservedGeneration,
@@ -390,23 +381,23 @@ func (r *DynamoGraphDeploymentRequestReconciler) Reconcile(ctx context.Context,
 	}
 	// State machine: handle different states
 	switch dgdr.Status.State {
-	case DGDRStateEmpty:
+	case nvidiacomv1alpha1.DGDRStateInitializing, "":
 		return r.handleInitialState(ctx, dgdr)
-	case DGDRStatePending:
+	case nvidiacomv1alpha1.DGDRStatePending:
 		return r.handlePendingState(ctx, dgdr)
-	case DGDRStateProfiling:
+	case nvidiacomv1alpha1.DGDRStateProfiling:
 		return r.handleProfilingState(ctx, dgdr)
-	case DGDRStateDeploying:
+	case nvidiacomv1alpha1.DGDRStateDeploying:
 		return r.handleDeployingState(ctx, dgdr)
-	case DGDRStateReady:
+	case nvidiacomv1alpha1.DGDRStateReady:
 		return r.handleReadyState(ctx, dgdr)
-	case DGDRStateDeploymentDeleted:
+	case nvidiacomv1alpha1.DGDRStateDeploymentDeleted:
 		return r.handleDeploymentDeletedState(ctx, dgdr)
-	case DGDRStateFailed:
+	case nvidiacomv1alpha1.DGDRStateFailed:
 		return r.handleFailedState(ctx, dgdr)
 	default:
 		logger.Info("Unknown state", "state", dgdr.Status.State)
-		return r.updateStateAndRequeue(ctx, dgdr, DGDRStateFailed, MessageInvalidState)
+		return r.updateStateAndRequeue(ctx, dgdr, nvidiacomv1alpha1.DGDRStateFailed, MessageInvalidState)
 	}
 }
 
@@ -418,7 +409,7 @@ func (r *DynamoGraphDeploymentRequestReconciler) handleInitialState(ctx context.
 	// Validate the spec
 	if err := r.validateSpec(ctx, dgdr); err != nil {
 		r.Recorder.Event(dgdr, corev1.EventTypeWarning, EventReasonValidationFailed, err.Error())
-		return r.updateStateWithCondition(ctx, dgdr, DGDRStateFailed, ConditionTypeValidation, metav1.ConditionFalse, EventReasonValidationFailed, err.Error())
+		return r.updateStateWithCondition(ctx, dgdr, nvidiacomv1alpha1.DGDRStateFailed, ConditionTypeValidation, metav1.ConditionFalse, EventReasonValidationFailed, err.Error())
 	}
 
 	// Set observedGeneration to track the spec we're processing
@@ -429,7 +420,7 @@ func (r *DynamoGraphDeploymentRequestReconciler) handleInitialState(ctx context.
 
 	// Initialize status
 	r.Recorder.Event(dgdr, corev1.EventTypeNormal, EventReasonInitialized, MessageInitialized)
-	return r.updateStateAndRequeue(ctx, dgdr, DGDRStatePending, MessageInitialized)
+	return r.updateStateAndRequeue(ctx, dgdr, nvidiacomv1alpha1.DGDRStatePending, MessageInitialized)
 }
 
 // handlePendingState starts the profiling process
@@ -440,7 +431,7 @@ func (r *DynamoGraphDeploymentRequestReconciler) handlePendingState(ctx context.
 	// Create profiling job (online or AIC)
 	if err := r.createProfilingJob(ctx, dgdr); err != nil {
 		r.Recorder.Event(dgdr, corev1.EventTypeWarning, EventReasonProfilingJobFailed, err.Error())
-		return r.updateStateWithCondition(ctx, dgdr, DGDRStateFailed, ConditionTypeProfiling, metav1.ConditionFalse, MessageJobCreationFailed, err.Error())
+		return r.updateStateWithCondition(ctx, dgdr, nvidiacomv1alpha1.DGDRStateFailed, ConditionTypeProfiling, metav1.ConditionFalse, MessageJobCreationFailed, err.Error())
 	}
 
 	// Record event with appropriate message
@@ -451,7 +442,7 @@ func (r *DynamoGraphDeploymentRequestReconciler) handlePendingState(ctx context.
 	}
 
 	// Update to Profiling state with Running status
-	return r.updateStateWithCondition(ctx, dgdr, DGDRStateProfiling, ConditionTypeProfiling, metav1.ConditionFalse, "ProfilingRunning", MessageProfilingInProgress)
+	return r.updateStateWithCondition(ctx, dgdr, nvidiacomv1alpha1.DGDRStateProfiling, ConditionTypeProfiling, metav1.ConditionFalse, "ProfilingRunning", MessageProfilingInProgress)
 }
 
 // handleProfilingState monitors profiling progress and generates spec when complete
@@ -465,7 +456,7 @@ func (r *DynamoGraphDeploymentRequestReconciler) handleProfilingState(ctx contex
 	if err != nil {
 		r.Recorder.Event(dgdr, corev1.EventTypeWarning, MessageProfilingCheckFailed, err.Error())
 		// Job failed - transition to Failed state
-		return r.updateStateWithCondition(ctx, dgdr, DGDRStateFailed, ConditionTypeProfiling, metav1.ConditionFalse, "ProfilingFailed", err.Error())
+		return r.updateStateWithCondition(ctx, dgdr, nvidiacomv1alpha1.DGDRStateFailed, ConditionTypeProfiling, metav1.ConditionFalse, "ProfilingFailed", err.Error())
 	}
 
 	if !completed {
@@ -486,7 +477,7 @@ func (r *DynamoGraphDeploymentRequestReconciler) handleProfilingState(ctx contex
 	// Retrieve profiling results and generate spec
 	if err := r.generateDGDSpec(ctx, dgdr); err != nil {
 		r.Recorder.Event(dgdr, corev1.EventTypeWarning, MessageGenerationFailed, err.Error())
-		return r.updateStateWithCondition(ctx, dgdr, DGDRStateFailed, ConditionTypeSpecGenerated, metav1.ConditionFalse, MessageGenerationFailed, err.Error())
+		return r.updateStateWithCondition(ctx, dgdr, nvidiacomv1alpha1.DGDRStateFailed, ConditionTypeSpecGenerated, metav1.ConditionFalse, MessageGenerationFailed, err.Error())
 	}
 
 	// Record spec generation event
@@ -508,11 +499,11 @@ func (r *DynamoGraphDeploymentRequestReconciler) handleProfilingState(ctx contex
 	// If autoApply is enabled, transition to Deploying state
 	if dgdr.Spec.AutoApply {
 		logger.Info("AutoApply enabled, transitioning to Deploying state")
-		return r.updateStateWithCondition(ctx, dgdr, DGDRStateDeploying, ConditionTypeSpecGenerated, metav1.ConditionTrue, EventReasonSpecGenerated, MessageSpecGenerated)
+		return r.updateStateWithCondition(ctx, dgdr, nvidiacomv1alpha1.DGDRStateDeploying, ConditionTypeSpecGenerated, metav1.ConditionTrue, EventReasonSpecGenerated, MessageSpecGenerated)
 	}
 
 	// Otherwise, transition to Ready state
-	return r.updateStateWithCondition(ctx, dgdr, DGDRStateReady, ConditionTypeSpecGenerated, metav1.ConditionTrue, EventReasonSpecGenerated, MessageSpecAvailable)
+	return r.updateStateWithCondition(ctx, dgdr, nvidiacomv1alpha1.DGDRStateReady, ConditionTypeSpecGenerated, metav1.ConditionTrue, EventReasonSpecGenerated, MessageSpecAvailable)
 }
 
 // handleReadyState handles DGDR in Ready state
@@ -549,7 +540,7 @@ func (r *DynamoGraphDeploymentRequestReconciler) handleReadyState(ctx context.Co
 		logger.Info("DGD degraded, transitioning back to Deploying",
 			"dgdState", dgd.Status.State)
 
-		dgdr.Status.State = DGDRStateDeploying
+		dgdr.Status.State = nvidiacomv1alpha1.DGDRStateDeploying
 
 		r.Recorder.Event(dgdr, corev1.EventTypeWarning, EventReasonDeploymentDegraded,
 			fmt.Sprintf(MessageDeploymentDegraded, dgd.Name, string(dgd.Status.State)))
@@ -573,7 +564,7 @@ func (r *DynamoGraphDeploymentRequestReconciler) handleDeployingState(ctx contex
 	if !dgdr.Spec.AutoApply {
 		// Shouldn't be in this state without autoApply
 		logger.Info("AutoApply not enabled, transitioning to Ready")
-		dgdr.Status.State = DGDRStateReady
+		dgdr.Status.State = nvidiacomv1alpha1.DGDRStateReady
 		return ctrl.Result{}, r.Status().Update(ctx, dgdr)
 	}
 
@@ -604,7 +595,7 @@ func (r *DynamoGraphDeploymentRequestReconciler) handleDeployingState(ctx contex
 	// Check if DGD is Ready
 	if dgd.Status.State == nvidiacomv1alpha1.DGDStateSuccessful {
 		logger.Info("DGD is Ready, transitioning to Ready state")
-		dgdr.Status.State = DGDRStateReady
+		dgdr.Status.State = nvidiacomv1alpha1.DGDRStateReady
 
 		r.Recorder.Event(dgdr, corev1.EventTypeNormal, EventReasonDeploymentReady,
 			fmt.Sprintf(MessageDeploymentReady, dgd.Name))
@@ -632,7 +623,7 @@ func (r *DynamoGraphDeploymentRequestReconciler) handleDGDDeleted(ctx context.Co
 	logger := log.FromContext(ctx)
 	logger.Info("DGD was deleted by user, transitioning to DeploymentDeleted state")
 
-	dgdr.Status.State = DGDRStateDeploymentDeleted
+	dgdr.Status.State = nvidiacomv1alpha1.DGDRStateDeploymentDeleted
 
 	r.Recorder.Event(dgdr, corev1.EventTypeWarning, EventReasonDeploymentDeleted,
 		fmt.Sprintf(MessageDeploymentDeleted, dgdr.Status.Deployment.Name))
@@ -1732,7 +1723,7 @@ func (r *DynamoGraphDeploymentRequestReconciler) extractDGDFromYAML(yamlContent
 }
 
 // updateStateAndRequeue updates the DGDR state and requeues
-func (r *DynamoGraphDeploymentRequestReconciler) updateStateAndRequeue(ctx context.Context, dgdr *nvidiacomv1alpha1.DynamoGraphDeploymentRequest, state, _ string) (ctrl.Result, error) {
+func (r *DynamoGraphDeploymentRequestReconciler) updateStateAndRequeue(ctx context.Context, dgdr *nvidiacomv1alpha1.DynamoGraphDeploymentRequest, state nvidiacomv1alpha1.DGDRState, _ string) (ctrl.Result, error) {
 	dgdr.Status.State = state
 	if err := r.Status().Update(ctx, dgdr); err != nil {
 		return ctrl.Result{}, err
@@ -1744,7 +1735,7 @@ func (r *DynamoGraphDeploymentRequestReconciler) updateStateAndRequeue(ctx conte
 func (r *DynamoGraphDeploymentRequestReconciler) updateStateWithCondition(
 	ctx context.Context,
 	dgdr *nvidiacomv1alpha1.DynamoGraphDeploymentRequest,
-	state string,
+	state nvidiacomv1alpha1.DGDRState,
 	conditionType string,
 	status metav1.ConditionStatus,
 	reason string,

deploy/operator/internal/controller/dynamographdeploymentrequest_controller_test.go

@@ -140,11 +140,11 @@ var _ = Describe("DynamoGraphDeploymentRequest Controller", func() {
 			Expect(err).NotTo(HaveOccurred())
 
 			// Check status
-			Eventually(func() string {
+			Eventually(func() nvidiacomv1alpha1.DGDRState {
 				var updated nvidiacomv1alpha1.DynamoGraphDeploymentRequest
 				_ = k8sClient.Get(ctx, types.NamespacedName{Name: dgdrName, Namespace: namespace}, &updated)
 				return updated.Status.State
-			}, timeout, interval).Should(Equal(DGDRStatePending))
+			}, timeout, interval).Should(Equal(nvidiacomv1alpha1.DGDRStatePending))
 
 			// Verify observedGeneration is set
 			var updated nvidiacomv1alpha1.DynamoGraphDeploymentRequest
@@ -190,11 +190,11 @@ var _ = Describe("DynamoGraphDeploymentRequest Controller", func() {
 			Expect(err).NotTo(HaveOccurred())
 
 			// Check status transitions to Pending (not Failed)
-			Eventually(func() string {
+			Eventually(func() nvidiacomv1alpha1.DGDRState {
 				var updated nvidiacomv1alpha1.DynamoGraphDeploymentRequest
 				_ = k8sClient.Get(ctx, types.NamespacedName{Name: dgdrName, Namespace: namespace}, &updated)
 				return updated.Status.State
-			}, timeout, interval).Should(Equal(DGDRStatePending))
+			}, timeout, interval).Should(Equal(nvidiacomv1alpha1.DGDRStatePending))
 		})
 	})
 
@@ -420,7 +420,7 @@ var _ = Describe("DynamoGraphDeploymentRequest Controller", func() {
 			defer func() { _ = k8sClient.Delete(ctx, dgdr) }()
 
 			// Update status to Profiling using Status subresource
-			dgdr.Status.State = DGDRStateProfiling
+			dgdr.Status.State = nvidiacomv1alpha1.DGDRStateProfiling
 			Expect(k8sClient.Status().Update(ctx, dgdr)).Should(Succeed())
 
 			// Create completed profiling job
@@ -495,7 +495,7 @@ spec:
 			Expect(updated.Status.GeneratedDeployment).NotTo(BeNil())
 
 			// Verify state transitioned to Ready (since autoApply is false by default)
-			Expect(updated.Status.State).Should(Equal(DGDRStateReady))
+			Expect(updated.Status.State).Should(Equal(nvidiacomv1alpha1.DGDRStateReady))
 		})
 	})
 
@@ -535,7 +535,7 @@ spec:
 			defer func() { _ = k8sClient.Delete(ctx, dgdr) }()
 
 			// Update status to Profiling using Status subresource
-			dgdr.Status.State = DGDRStateProfiling
+			dgdr.Status.State = nvidiacomv1alpha1.DGDRStateProfiling
 			Expect(k8sClient.Status().Update(ctx, dgdr)).Should(Succeed())
 
 			// Create completed profiling job
@@ -605,7 +605,7 @@ spec:
 			// Get updated DGDR and check state is Deploying
 			var updated nvidiacomv1alpha1.DynamoGraphDeploymentRequest
 			Expect(k8sClient.Get(ctx, types.NamespacedName{Name: dgdrName, Namespace: namespace}, &updated)).Should(Succeed())
-			Expect(updated.Status.State).Should(Equal(DGDRStateDeploying))
+			Expect(updated.Status.State).Should(Equal(nvidiacomv1alpha1.DGDRStateDeploying))
 
 			// Reconcile again to create DGD
 			_, err = reconciler.Reconcile(ctx, reconcile.Request{
@@ -676,7 +676,7 @@ spec:
 			observedGeneration := current.Status.ObservedGeneration
 
 			// Manually set state to Profiling to simulate in-progress profiling
-			current.Status.State = DGDRStateProfiling
+			current.Status.State = nvidiacomv1alpha1.DGDRStateProfiling
 			Expect(k8sClient.Status().Update(ctx, &current)).Should(Succeed())
 
 			// Try to modify spec
@@ -698,7 +698,7 @@ spec:
 			Expect(k8sClient.Get(ctx, types.NamespacedName{Name: dgdrName, Namespace: namespace}, &current)).Should(Succeed())
 			Expect(current.Generation).Should(BeNumerically(">", initialGeneration))
 			Expect(current.Status.ObservedGeneration).Should(Equal(observedGeneration))
-			Expect(current.Status.State).Should(Equal(DGDRStateProfiling)) // State unchanged
+			Expect(current.Status.State).Should(Equal(nvidiacomv1alpha1.DGDRStateProfiling)) // State unchanged
 
 			// Verify event was recorded
 			Eventually(func() bool {
@@ -748,7 +748,7 @@ spec:
 			defer func() { _ = k8sClient.Delete(ctx, dgdr) }()
 
 			// Update status to Ready with Deployment info using Status subresource
-			dgdr.Status.State = DGDRStateReady
+			dgdr.Status.State = nvidiacomv1alpha1.DGDRStateReady
 			dgdr.Status.Deployment = &nvidiacomv1alpha1.DeploymentStatus{
 				Name:      "test-dgd-to-delete",
 				Namespace: namespace,
@@ -766,7 +766,7 @@ spec:
 			// Get updated DGDR and check state transitioned to DeploymentDeleted
 			var updated nvidiacomv1alpha1.DynamoGraphDeploymentRequest
 			Expect(k8sClient.Get(ctx, types.NamespacedName{Name: dgdrName, Namespace: namespace}, &updated)).Should(Succeed())
-			Expect(updated.Status.State).Should(Equal(DGDRStateDeploymentDeleted))
+			Expect(updated.Status.State).Should(Equal(nvidiacomv1alpha1.DGDRStateDeploymentDeleted))
 		})
 	})
 })
@@ -1241,7 +1241,7 @@ var _ = Describe("DGDR Error Handling", func() {
 			defer func() { _ = k8sClient.Delete(ctx, dgdr) }()
 
 			// Set status to Profiling
-			dgdr.Status.State = DGDRStateProfiling
+			dgdr.Status.State = nvidiacomv1alpha1.DGDRStateProfiling
 			Expect(k8sClient.Status().Update(ctx, dgdr)).Should(Succeed())
 
 			// Create failed job
@@ -1323,7 +1323,7 @@ var _ = Describe("DGDR Error Handling", func() {
 			// Verify DGDR transitioned to Failed state
 			var updated nvidiacomv1alpha1.DynamoGraphDeploymentRequest
 			Expect(k8sClient.Get(ctx, types.NamespacedName{Name: dgdrName, Namespace: namespace}, &updated)).Should(Succeed())
-			Expect(updated.Status.State).Should(Equal(DGDRStateFailed))
+			Expect(updated.Status.State).Should(Equal(nvidiacomv1alpha1.DGDRStateFailed))
 
 			// Verify error condition contains detailed error
 			condition := meta.FindStatusCondition(updated.Status.Conditions, ConditionTypeProfiling)
@@ -1560,7 +1560,7 @@ spec:
 			// Should transition to Pending (validation passed)
 			var updated nvidiacomv1alpha1.DynamoGraphDeploymentRequest
 			_ = k8sClient.Get(ctx, types.NamespacedName{Name: dgdrName, Namespace: namespace}, &updated)
-			Expect(updated.Status.State).Should(Equal(DGDRStatePending))
+			Expect(updated.Status.State).Should(Equal(nvidiacomv1alpha1.DGDRStatePending))
 		})
 
 		It("Should respect manual hardware config over GPU discovery", func() {
@@ -1623,7 +1623,7 @@ spec:
 			// Should transition to Pending (validation passed with manual config)
 			var updated nvidiacomv1alpha1.DynamoGraphDeploymentRequest
 			_ = k8sClient.Get(ctx, types.NamespacedName{Name: dgdrName, Namespace: namespace}, &updated)
-			Expect(updated.Status.State).Should(Equal(DGDRStatePending))
+			Expect(updated.Status.State).Should(Equal(nvidiacomv1alpha1.DGDRStatePending))
 		})
 
 		It("Should succeed with GPU discovery when cluster has GPU nodes", func() {
@@ -1680,7 +1680,7 @@ spec:
 			// Should transition to Pending
 			var updated nvidiacomv1alpha1.DynamoGraphDeploymentRequest
 			_ = k8sClient.Get(ctx, types.NamespacedName{Name: dgdrName, Namespace: namespace}, &updated)
-			Expect(updated.Status.State).Should(Equal(DGDRStatePending))
+			Expect(updated.Status.State).Should(Equal(nvidiacomv1alpha1.DGDRStatePending))
 		})
 
 		It("Should pass validation with explicit GPU ranges without GPU discovery", func() {
@@ -1731,7 +1731,7 @@ spec:
 			// Should transition to Pending
 			var updated nvidiacomv1alpha1.DynamoGraphDeploymentRequest
 			_ = k8sClient.Get(ctx, types.NamespacedName{Name: dgdrName, Namespace: namespace}, &updated)
-			Expect(updated.Status.State).Should(Equal(DGDRStatePending))
+			Expect(updated.Status.State).Should(Equal(nvidiacomv1alpha1.DGDRStatePending))
 		})
 
 		It("Should use GPU discovery with heterogeneous nodes (picks best)", func() {
@@ -1803,7 +1803,7 @@ spec:
 			// Should transition to Pending (using H100 config)
 			var updated nvidiacomv1alpha1.DynamoGraphDeploymentRequest
 			_ = k8sClient.Get(ctx, types.NamespacedName{Name: dgdrName, Namespace: namespace}, &updated)
-			Expect(updated.Status.State).Should(Equal(DGDRStatePending))
+			Expect(updated.Status.State).Should(Equal(nvidiacomv1alpha1.DGDRStatePending))
 		})
 	})
 })

docs/pages/kubernetes/api-reference.md

@@ -112,6 +112,29 @@ _Appears in:_
 | `key` _string_ | Key in the ConfigMap to select. If not specified, defaults to "disagg.yaml". | disagg.yaml |  |
 
 
+#### DGDRState
+
+_Underlying type:_ _string_
+
+
+
+_Validation:_
+- Enum: [Initializing Pending Profiling Deploying Ready DeploymentDeleted Failed]
+
+_Appears in:_
+- [DynamoGraphDeploymentRequestStatus](#dynamographdeploymentrequeststatus)
+
+| Field | Description |
+| --- | --- |
+| `Initializing` |  |
+| `Pending` |  |
+| `Profiling` |  |
+| `Deploying` |  |
+| `Ready` |  |
+| `DeploymentDeleted` |  |
+| `Failed` |  |
+
+
 #### DGDState
 
 _Underlying type:_ _string_
@@ -442,7 +465,7 @@ It serves as the primary interface for users to request model deployments with
 specific performance and resource constraints, enabling SLA-driven deployments.
 
 Lifecycle:
- 1. Initial → Pending: Validates spec and prepares for profiling
+ 1. Initializing → Pending: Validates spec and prepares for profiling
  2. Pending → Profiling: Creates and runs profiling job (online or AIC)
  3. Profiling → Ready/Deploying: Generates DGD spec after profiling completes
  4. Deploying → Ready: When autoApply=true, monitors DGD until Ready
@@ -503,7 +526,7 @@ _Appears in:_
 
 | Field | Description | Default | Validation |
 | --- | --- | --- | --- |
-| `state` _string_ | State is a high-level textual status of the deployment request lifecycle.<br />Possible values: "", "Pending", "Profiling", "Deploying", "Ready", "DeploymentDeleted", "Failed"<br />Empty string ("") represents the initial state before initialization. |  |  |
+| `state` _[DGDRState](#dgdrstate)_ | State is a high-level textual status of the deployment request lifecycle. | Initializing | Enum: [Initializing Pending Profiling Deploying Ready DeploymentDeleted Failed] <br /> |
 | `backend` _string_ | Backend is extracted from profilingConfig.config.engine.backend for display purposes.<br />This field is populated by the controller and shown in kubectl output. |  | Optional: \{\} <br /> |
 | `observedGeneration` _integer_ | ObservedGeneration reflects the generation of the most recently observed spec.<br />Used to detect spec changes and enforce immutability after profiling starts. |  |  |
 | `conditions` _[Condition](https://kubernetes.io/docs/reference/generated/kubernetes-api/v1.28/#condition-v1-meta) array_ | Conditions contains the latest observed conditions of the deployment request.<br />Standard condition types include: Validation, Profiling, SpecGenerated, DeploymentReady.<br />Conditions are merged by type on patch updates. |  |  |