Merge remote-tracking branch 'upstream/main' into issue-123

README.md

@@ -132,6 +132,31 @@ If these parameters are unspecified, the [default image and tag](config/base/par
 If you'd like to change the service image/tag after deploying the operator, simply change the parameters in the KFDef. Any
 TrustyAI service deployed subsequently will use the new image and tag. 
 
+### `TrustyAIService` Status Updates
+
+The `TrustyAIService` custom resource tracks the availability of `InferenceServices` and `PersistentVolumeClaims (PVCs)` 
+through its `status` field. Below are the status types and reasons that are available:
+
+#### `InferenceService` Status
+
+| Status Type                   | Status Reason                     | Description                       |
+|-------------------------------|-----------------------------------|-----------------------------------|
+| `InferenceServicesPresent`    | `InferenceServicesNotFound`       | InferenceServices were not found. |
+| `InferenceServicesPresent`    | `InferenceServicesFound`          | InferenceServices were found.     |
+
+#### `PersistentVolumeClaim` (PVCs) Status
+
+| Status Type      | Status Reason   | Description                        |
+|------------------|-----------------|------------------------------------|
+| `PVCAvailable`   | `PVCNotFound`   | `PersistentVolumeClaim` not found.  |
+| `PVCAvailable`   | `PVCFound`      | `PersistentVolumeClaim` found.      |
+
+
+#### Status Behavior
+
+- If a PVC is not available, the `Ready` status of `TrustyAIService` will be set to `False`.
+- However, if `InferenceServices` are not found, the `Ready` status of `TrustyAIService` will not be affected, _i.e._, it is `Ready` by all other conditions, it will remain so.
+
 ## Contributing
 
 Please see the [CONTRIBUTING.md](./CONTRIBUTING.md) file for more details on how to contribute to this project.

api/v1alpha1/trustyaiservice_types.go

@@ -89,3 +89,27 @@ type TrustyAIServiceList struct {
 func init() {
 	SchemeBuilder.Register(&TrustyAIService{}, &TrustyAIServiceList{})
 }
+
+// SetStatus sets the status of the TrustyAIService
+func (t *TrustyAIService) SetStatus(condType, reason, message string, status corev1.ConditionStatus) {
+	now := metav1.Now()
+	condition := Condition{
+		Type:               condType,
+		Status:             status,
+		Reason:             reason,
+		Message:            message,
+		LastTransitionTime: now,
+	}
+	// Replace or append condition
+	found := false
+	for i, cond := range t.Status.Conditions {
+		if cond.Type == condType {
+			t.Status.Conditions[i] = condition
+			found = true
+			break
+		}
+	}
+	if !found {
+		t.Status.Conditions = append(t.Status.Conditions, condition)
+	}
+}

controllers/constants.go

@@ -14,3 +14,20 @@ const (
 	volumeMountName      = "volume"
 	defaultRequeueDelay  = time.Minute
 )
+
+// Status types
+const (
+	StatusTypeInferenceServicesPresent = "InferenceServicesPresent"
+	StatusTypePVCAvailable             = "PVCAvailable"
+	StatusTypeRouteAvailable           = "RouteAvailable"
+)
+
+// Status reasons
+const (
+	StatusReasonInferenceServicesNotFound = "InferenceServicesNotFound"
+	StatusReasonInferenceServicesFound    = "InferenceServicesFound"
+	StatusReasonPVCNotFound               = "PVCNotFound"
+	StatusReasonPVCFound                  = "PVCFound"
+	StatusReasonRouteNotFound             = "RouteNotFound"
+	StatusReasonRouteFound                = "RouteFound"
+)

controllers/inference_services.go

@@ -4,6 +4,7 @@ import (
 	"context"
 	"fmt"
 	kservev1beta1 "github.com/kserve/kserve/pkg/apis/serving/v1beta1"
+	trustyaiopendatahubiov1alpha1 "github.com/trustyai-explainability/trustyai-service-operator/api/v1alpha1"
 	appsv1 "k8s.io/api/apps/v1"
 	corev1 "k8s.io/api/core/v1"
 	"sigs.k8s.io/controller-runtime/pkg/client"
@@ -127,14 +128,23 @@ func generateEnvVarValue(currentValue, newValue string, remove bool) string {
 	return currentValue
 }
 
-func (r *TrustyAIServiceReconciler) handleInferenceServices(ctx context.Context, namespace string, labelKey string, labelValue string, envVarName string, crName string, remove bool) (bool, error) {
+func (r *TrustyAIServiceReconciler) handleInferenceServices(ctx context.Context, instance *trustyaiopendatahubiov1alpha1.TrustyAIService, namespace string, labelKey string, labelValue string, envVarName string, crName string, remove bool) (bool, error) {
 	var inferenceServices kservev1beta1.InferenceServiceList
 
 	if err := r.List(ctx, &inferenceServices, client.InNamespace(namespace)); err != nil {
 		log.FromContext(ctx).Error(err, "Could not list InferenceService objects.")
 		return false, err
 	}
 
+	if len(inferenceServices.Items) == 0 {
+		_, updateErr := r.updateStatus(ctx, instance, UpdateInferenceServiceNotPresent)
+		if updateErr != nil {
+			log.FromContext(ctx).Error(updateErr, "Could not update status for InferenceService not present")
+			return false, updateErr
+		}
+		return true, nil
+	}
+
 	for _, infService := range inferenceServices.Items {
 		annotations := infService.GetAnnotations()
 		// Check the annotation "serving.kserve.io/deploymentMode: ModelMesh"
@@ -152,6 +162,13 @@ func (r *TrustyAIServiceReconciler) handleInferenceServices(ctx context.Context,
 			}
 		}
 	}
+
+	instance.SetStatus("InferenceServicesPresent", "InferenceServicesFound", "InferenceServices found", corev1.ConditionTrue)
+	_, updateErr := r.updateStatus(ctx, instance, UpdateInferenceServicePresent)
+	if updateErr != nil {
+		log.FromContext(ctx).Error(updateErr, "Could not update status for InferenceService present")
+		return false, updateErr
+	}
 	return true, nil
 }
 

controllers/statuses.go

@@ -0,0 +1,60 @@
+package controllers
+
+import (
+	"context"
+	trustyaiopendatahubiov1alpha1 "github.com/trustyai-explainability/trustyai-service-operator/api/v1alpha1"
+	v1 "k8s.io/api/core/v1"
+	"k8s.io/client-go/util/retry"
+	"sigs.k8s.io/controller-runtime/pkg/client"
+	"sigs.k8s.io/controller-runtime/pkg/log"
+)
+
+func (r *TrustyAIServiceReconciler) updateStatus(ctx context.Context, original *trustyaiopendatahubiov1alpha1.TrustyAIService, update func(saved *trustyaiopendatahubiov1alpha1.TrustyAIService),
+) (*trustyaiopendatahubiov1alpha1.TrustyAIService, error) {
+	saved := &trustyaiopendatahubiov1alpha1.TrustyAIService{}
+	err := retry.RetryOnConflict(retry.DefaultRetry, func() error {
+		err := r.Client.Get(ctx, client.ObjectKeyFromObject(original), saved)
+		if err != nil {
+			return err
+		}
+		// update status here
+		update(saved)
+
+		// Try to update
+		err = r.Client.Status().Update(ctx, saved)
+		return err
+	})
+	if err != nil {
+		log.FromContext(ctx).Error(err, "Failed to update TrustyAIService status")
+	}
+	return saved, err
+}
+
+func UpdateInferenceServiceNotPresent(saved *trustyaiopendatahubiov1alpha1.TrustyAIService) {
+	saved.SetStatus(StatusTypeInferenceServicesPresent, StatusReasonInferenceServicesNotFound, "InferenceServices not found", v1.ConditionFalse)
+	saved.Status.Phase = "Not Ready"
+	saved.Status.Ready = v1.ConditionFalse
+
+}
+
+func UpdateInferenceServicePresent(saved *trustyaiopendatahubiov1alpha1.TrustyAIService) {
+	saved.SetStatus(StatusTypeInferenceServicesPresent, StatusReasonInferenceServicesFound, "InferenceServices found", v1.ConditionTrue)
+}
+
+func UpdatePVCNotAvailable(saved *trustyaiopendatahubiov1alpha1.TrustyAIService) {
+	saved.SetStatus(StatusTypePVCAvailable, StatusReasonPVCNotFound, "PersistentVolumeClaim not found", v1.ConditionFalse)
+	saved.Status.Phase = "Not Ready"
+	saved.Status.Ready = v1.ConditionFalse
+}
+
+func UpdatePVCAvailable(saved *trustyaiopendatahubiov1alpha1.TrustyAIService) {
+	saved.SetStatus(StatusTypePVCAvailable, StatusReasonPVCFound, "PersistentVolumeClaim found", v1.ConditionTrue)
+}
+
+func UpdateRouteAvailable(saved *trustyaiopendatahubiov1alpha1.TrustyAIService) {
+	saved.SetStatus(StatusTypeRouteAvailable, StatusReasonRouteFound, "Route found", v1.ConditionTrue)
+}
+
+func UpdateRouteNotAvailable(saved *trustyaiopendatahubiov1alpha1.TrustyAIService) {
+	saved.SetStatus(StatusTypeRouteAvailable, StatusReasonRouteNotFound, "Route not found", v1.ConditionFalse)
+}