feat: add RayCluster.status.metadata field

as map[string]string. First entry will be head node IP.

apiserver/pkg/model/converter.go

@@ -55,6 +55,8 @@ func FromCrdToApiCluster(cluster *v1alpha1.RayCluster, events []v1.Event) *api.C
 	for name, port := range cluster.Status.Endpoints {
 		pbCluster.ServiceEndpoint[name] = port
 	}
+
+	pbCluster.Metadata = cluster.Status.Metadata
 	return pbCluster
 }
 

proto/cluster.proto

@@ -161,6 +161,9 @@ message Cluster {
 
   // Output. The service endpoint of the cluster
   map<string, string> service_endpoint = 11;
+
+  // Output. Various metadata about the cluster.
+  map<string, string> metadata = 12;
 }
 
 message ClusterSpec {

proto/swagger/cluster.swagger.json

@@ -299,6 +299,13 @@
             "type": "string"
           },
           "title": "Output. The service endpoint of the cluster"
+        },
+        "metadata": {
+          "type": "object",
+          "additionalProperties": {
+            "type": "string"
+          },
+          "description": "Output. Various metadata about the cluster."
         }
       }
     },

ray-operator/apis/ray/v1alpha1/raycluster_types.go

@@ -112,6 +112,8 @@ type RayClusterStatus struct {
 	LastUpdateTime *metav1.Time `json:"lastUpdateTime,omitempty"`
 	// Service Endpoints
 	Endpoints map[string]string `json:"endpoints,omitempty"`
+	// Metadata
+	Metadata map[string]string `json:"metadata,omitempty"`
 }
 
 // RayNodeType  the type of a ray node: head/worker

ray-operator/config/crd/bases/ray.io_rayclusters.yaml

@@ -10992,6 +10992,11 @@ spec:
                   type: string
                 description: Service Endpoints
                 type: object
+              metadata:
+                additionalProperties:
+                  type: string
+                description: Metadata
+                type: object
               lastUpdateTime:
                 description: LastUpdateTime indicates last update timestamp for this
                   cluster status.

ray-operator/config/crd/bases/ray.io_rayservices.yaml

@@ -11600,6 +11600,11 @@ spec:
                           type: string
                         description: Service Endpoints
                         type: object
+                      metadata:
+                        additionalProperties:
+                          type: string
+                        description: Metadata
+                        type: object
                       lastUpdateTime:
                         description: LastUpdateTime indicates last update timestamp
                           for this cluster status.
@@ -11699,6 +11704,11 @@ spec:
                           type: string
                         description: Service Endpoints
                         type: object
+                      metadata:
+                        additionalProperties:
+                          type: string
+                        description: Metadata
+                        type: object
                       lastUpdateTime:
                         description: LastUpdateTime indicates last update timestamp
                           for this cluster status.

ray-operator/controllers/ray/common/constant.go

@@ -88,6 +88,8 @@ const (
 	// Ray health check related configurations
 	RayAgentRayletHealthPath  = "api/local_raylet_healthz"
 	RayDashboardGCSHealthPath = "api/gcs_healthz"
+	// Cluster metadata keys
+	HeadNodeIpKey = "head_node_ip"
 )
 
 type ServiceType string

ray-operator/controllers/ray/raycluster_controller.go

@@ -778,6 +778,10 @@ func (r *RayClusterReconciler) updateStatus(instance *rayiov1alpha1.RayCluster)
 		return err
 	}
 
+	if err := r.updateMetadata(instance); err != nil {
+		return err
+	}
+
 	timeNow := metav1.Now()
 	instance.Status.LastUpdateTime = &timeNow
 	if err := r.Status().Update(context.Background(), instance); err != nil {
@@ -955,3 +959,21 @@ func (r *RayClusterReconciler) updateClusterState(instance *rayiov1alpha1.RayClu
 	instance.Status.State = clusterState
 	return r.Status().Update(context.Background(), instance)
 }
+
+func (r *RayClusterReconciler) updateMetadata(instance *rayiov1alpha1.RayCluster) error {
+	runtimePods := corev1.PodList{}
+	filterLabels := client.MatchingLabels{common.RayClusterLabelKey: instance.Name, common.RayNodeTypeLabelKey: string(rayiov1alpha1.HeadNode)}
+	if err := r.List(context.TODO(), &runtimePods, client.InNamespace(instance.Namespace), filterLabels); err != nil {
+		return err
+	}
+	if len(runtimePods.Items) < 1 {
+		r.Log.Info("Unable to find head node pod. Not updating status with head IP", "cluster name", instance.Name, "filter labels", filterLabels)
+	} else if len(runtimePods.Items) > 1 {
+		r.Log.Info("Found multiple head node pods. Not updating status with head IP", "cluster name", instance.Name, "filter labels", filterLabels)
+	}
+
+	instance.Status.Metadata = map[string]string{
+		common.HeadNodeIpKey: runtimePods.Items[0].Status.PodIP,
+	}
+	return nil
+}

ray-operator/controllers/ray/raycluster_controller_fake_test.go

@@ -98,6 +98,7 @@ func setupTest(t *testing.T) {
 			},
 			Status: corev1.PodStatus{
 				Phase: corev1.PodRunning,
+				PodIP: "123.123.123.123",
 			},
 		},
 		&corev1.Pod{
@@ -779,3 +780,26 @@ func TestUpdateEndpoints(t *testing.T) {
 	}
 	assert.Equal(t, expected, testRayCluster.Status.Endpoints, "RayCluster status endpoints not updated")
 }
+
+func TestUpdateMetadata(t *testing.T) {
+	setupTest(t)
+	defer tearDown(t)
+
+	fakeClient := clientFake.NewClientBuilder().WithRuntimeObjects(testPods...).Build()
+
+	testRayClusterReconciler := &RayClusterReconciler{
+		Client:   fakeClient,
+		Recorder: &record.FakeRecorder{},
+		Scheme:   scheme.Scheme,
+		Log:      ctrl.Log.WithName("controllers").WithName("RayCluster"),
+	}
+
+	if err := testRayClusterReconciler.updateMetadata(testRayCluster); err != nil {
+		t.Errorf("updateMetadata failed: %v", err)
+	}
+
+	expected := map[string]string{
+		"head_node_ip": "123.123.123.123",
+	}
+	assert.Equal(t, expected, testRayCluster.Status.Metadata, "RayCluster status.metadata not updated")
+}