From 13eb7b21714002bd42f8c6a57642d81651f3ed44 Mon Sep 17 00:00:00 2001
From: Chia-Wei <56123090+cchen777@users.noreply.github.com>
Date: Fri, 6 Sep 2024 00:35:49 +0800
Subject: [PATCH] [Feature] Display reconcile failures as events
 (ServiceAccount) (#2290)

* [Feature] Display reconcile failures as events on ray clusters

* reproduce

* clean up

* add more eventf for sa specific

* add eventf for failed to createHeadPod, this can happen when service account not available in namespace

* add woker pod create fail event

* reuse actionable msg

* add AutoscalerServiceAccountNotFound new K8sEventType
---
 ray-operator/controllers/ray/raycluster_controller.go | 11 +++++++----
 ray-operator/controllers/ray/utils/constant.go        |  5 +++--
 2 files changed, 10 insertions(+), 6 deletions(-)

diff --git a/ray-operator/controllers/ray/raycluster_controller.go b/ray-operator/controllers/ray/raycluster_controller.go
index c6df3b6262..6138befdc9 100644
--- a/ray-operator/controllers/ray/raycluster_controller.go
+++ b/ray-operator/controllers/ray/raycluster_controller.go
@@ -1375,10 +1375,13 @@ func (r *RayClusterReconciler) reconcileAutoscalerServiceAccount(ctx context.Con
 		// zero-downtime rolling updates when RayService is performed. See https://github.com/ray-project/kuberay/issues/1123
 		// for more details.
 		if instance.Spec.HeadGroupSpec.Template.Spec.ServiceAccountName == namespacedName.Name {
-			logger.Error(err, fmt.Sprintf(
-				"If users specify ServiceAccountName for the head Pod, they need to create a ServiceAccount themselves. "+
-					"However, ServiceAccount %s is not found. Please create one. "+
-					"See the PR description of https://github.com/ray-project/kuberay/pull/1128 for more details.", namespacedName.Name), "ServiceAccount", namespacedName)
+			actionableMessage := fmt.Sprintf("If users specify ServiceAccountName for the head Pod, they need to create a ServiceAccount themselves. "+
+				"However, ServiceAccount %s is not found. Please create one. See the PR description of https://github.com/ray-project/kuberay/pull/1128 for more details.", namespacedName.Name)
+
+			logger.Error(
+				err,
+				actionableMessage)
+			r.Recorder.Eventf(instance, corev1.EventTypeWarning, string(utils.AutoscalerServiceAccountNotFound), "Failed to reconcile RayCluster %s/%s. %s", instance.Namespace, instance.Name, actionableMessage)
 			return err
 		}
 
diff --git a/ray-operator/controllers/ray/utils/constant.go b/ray-operator/controllers/ray/utils/constant.go
index b0375968e4..a6fd2806c2 100644
--- a/ray-operator/controllers/ray/utils/constant.go
+++ b/ray-operator/controllers/ray/utils/constant.go
@@ -256,8 +256,9 @@ const (
 	FailedToCreateService K8sEventType = "FailedToCreateService"
 
 	// ServiceAccount event list
-	CreatedServiceAccount        K8sEventType = "CreatedServiceAccount"
-	FailedToCreateServiceAccount K8sEventType = "FailedToCreateServiceAccount"
+	CreatedServiceAccount            K8sEventType = "CreatedServiceAccount"
+	FailedToCreateServiceAccount     K8sEventType = "FailedToCreateServiceAccount"
+	AutoscalerServiceAccountNotFound K8sEventType = "AutoscalerServiceAccountNotFound"
 
 	// Role event list
 	CreatedRole        K8sEventType = "CreatedRole"