feat: Added support for enabling metrics in Feast Operator (#5317)

infra/feast-operator/api/v1alpha1/featurestore_types.go

@@ -515,6 +515,8 @@ type ServerConfigs struct {
 	// Allowed values: "debug", "info", "warning", "error", "critical".
 	// +kubebuilder:validation:Enum=debug;info;warning;error;critical
 	LogLevel *string `json:"logLevel,omitempty"`
+	// Metrics exposes Prometheus-compatible metrics for the Feast server when enabled.
+	Metrics *bool `json:"metrics,omitempty"`
 	// VolumeMounts defines the list of volumes that should be mounted into the feast container.
 	// This allows attaching persistent storage, config files, secrets, or other resources
 	// required by the Feast components. Ensure that each volume mount has a corresponding

infra/feast-operator/config/crd/bases/feast.dev_featurestores.yaml

@@ -1041,6 +1041,10 @@ spec:
                             - error
                             - critical
                             type: string
+                          metrics:
+                            description: Metrics exposes Prometheus-compatible metrics
+                              for the Feast server when enabled.
+                            type: boolean
                           nodeSelector:
                             additionalProperties:
                               type: string
@@ -1500,6 +1504,10 @@ spec:
                             - error
                             - critical
                             type: string
+                          metrics:
+                            description: Metrics exposes Prometheus-compatible metrics
+                              for the Feast server when enabled.
+                            type: boolean
                           nodeSelector:
                             additionalProperties:
                               type: string
@@ -1985,6 +1993,10 @@ spec:
                                 - error
                                 - critical
                                 type: string
+                              metrics:
+                                description: Metrics exposes Prometheus-compatible
+                                  metrics for the Feast server when enabled.
+                                type: boolean
                               nodeSelector:
                                 additionalProperties:
                                   type: string
@@ -2484,6 +2496,10 @@ spec:
                         - error
                         - critical
                         type: string
+                      metrics:
+                        description: Metrics exposes Prometheus-compatible metrics
+                          for the Feast server when enabled.
+                        type: boolean
                       nodeSelector:
                         additionalProperties:
                           type: string
@@ -5057,6 +5073,10 @@ spec:
                                 - error
                                 - critical
                                 type: string
+                              metrics:
+                                description: Metrics exposes Prometheus-compatible
+                                  metrics for the Feast server when enabled.
+                                type: boolean
                               nodeSelector:
                                 additionalProperties:
                                   type: string
@@ -5524,6 +5544,10 @@ spec:
                                 - error
                                 - critical
                                 type: string
+                              metrics:
+                                description: Metrics exposes Prometheus-compatible
+                                  metrics for the Feast server when enabled.
+                                type: boolean
                               nodeSelector:
                                 additionalProperties:
                                   type: string
@@ -6021,6 +6045,10 @@ spec:
                                     - error
                                     - critical
                                     type: string
+                                  metrics:
+                                    description: Metrics exposes Prometheus-compatible
+                                      metrics for the Feast server when enabled.
+                                    type: boolean
                                   nodeSelector:
                                     additionalProperties:
                                       type: string
@@ -6530,6 +6558,10 @@ spec:
                             - error
                             - critical
                             type: string
+                          metrics:
+                            description: Metrics exposes Prometheus-compatible metrics
+                              for the Feast server when enabled.
+                            type: boolean
                           nodeSelector:
                             additionalProperties:
                               type: string

infra/feast-operator/dist/install.yaml

@@ -1049,6 +1049,10 @@ spec:
                             - error
                             - critical
                             type: string
+                          metrics:
+                            description: Metrics exposes Prometheus-compatible metrics
+                              for the Feast server when enabled.
+                            type: boolean
                           nodeSelector:
                             additionalProperties:
                               type: string
@@ -1508,6 +1512,10 @@ spec:
                             - error
                             - critical
                             type: string
+                          metrics:
+                            description: Metrics exposes Prometheus-compatible metrics
+                              for the Feast server when enabled.
+                            type: boolean
                           nodeSelector:
                             additionalProperties:
                               type: string
@@ -1993,6 +2001,10 @@ spec:
                                 - error
                                 - critical
                                 type: string
+                              metrics:
+                                description: Metrics exposes Prometheus-compatible
+                                  metrics for the Feast server when enabled.
+                                type: boolean
                               nodeSelector:
                                 additionalProperties:
                                   type: string
@@ -2492,6 +2504,10 @@ spec:
                         - error
                         - critical
                         type: string
+                      metrics:
+                        description: Metrics exposes Prometheus-compatible metrics
+                          for the Feast server when enabled.
+                        type: boolean
                       nodeSelector:
                         additionalProperties:
                           type: string
@@ -5065,6 +5081,10 @@ spec:
                                 - error
                                 - critical
                                 type: string
+                              metrics:
+                                description: Metrics exposes Prometheus-compatible
+                                  metrics for the Feast server when enabled.
+                                type: boolean
                               nodeSelector:
                                 additionalProperties:
                                   type: string
@@ -5532,6 +5552,10 @@ spec:
                                 - error
                                 - critical
                                 type: string
+                              metrics:
+                                description: Metrics exposes Prometheus-compatible
+                                  metrics for the Feast server when enabled.
+                                type: boolean
                               nodeSelector:
                                 additionalProperties:
                                   type: string
@@ -6029,6 +6053,10 @@ spec:
                                     - error
                                     - critical
                                     type: string
+                                  metrics:
+                                    description: Metrics exposes Prometheus-compatible
+                                      metrics for the Feast server when enabled.
+                                    type: boolean
                                   nodeSelector:
                                     additionalProperties:
                                       type: string
@@ -6538,6 +6566,10 @@ spec:
                             - error
                             - critical
                             type: string
+                          metrics:
+                            description: Metrics exposes Prometheus-compatible metrics
+                              for the Feast server when enabled.
+                            type: boolean
                           nodeSelector:
                             additionalProperties:
                               type: string

infra/feast-operator/docs/api/markdown/ref.md

@@ -702,6 +702,7 @@ _Appears in:_
 | `tls` _[TlsConfigs](#tlsconfigs)_ |  |
 | `logLevel` _string_ | LogLevel sets the logging level for the server
 Allowed values: "debug", "info", "warning", "error", "critical". |
+| `metrics` _boolean_ | Metrics exposes Prometheus-compatible metrics for the Feast server when enabled. |
 | `volumeMounts` _[VolumeMount](https://kubernetes.io/docs/reference/generated/kubernetes-api/v1.30/#volumemount-v1-core) array_ | VolumeMounts defines the list of volumes that should be mounted into the feast container.
 This allows attaching persistent storage, config files, secrets, or other resources
 required by the Feast components. Ensure that each volume mount has a corresponding
@@ -765,6 +766,7 @@ _Appears in:_
 | `tls` _[TlsConfigs](#tlsconfigs)_ |  |
 | `logLevel` _string_ | LogLevel sets the logging level for the server
 Allowed values: "debug", "info", "warning", "error", "critical". |
+| `metrics` _boolean_ | Metrics exposes Prometheus-compatible metrics for the Feast server when enabled. |
 | `volumeMounts` _[VolumeMount](https://kubernetes.io/docs/reference/generated/kubernetes-api/v1.30/#volumemount-v1-core) array_ | VolumeMounts defines the list of volumes that should be mounted into the feast container.
 This allows attaching persistent storage, config files, secrets, or other resources
 required by the Feast components. Ensure that each volume mount has a corresponding

infra/feast-operator/internal/controller/services/services.go

@@ -470,6 +470,13 @@ func (feast *FeastServices) setContainer(containers *[]corev1.Container, feastTy
 				ContainerPort: getTargetPort(feastType, tls),
 				Protocol:      corev1.ProtocolTCP,
 			})
+			if feastType == OnlineFeastType && feast.isMetricsEnabled(feastType) {
+				container.Ports = append(container.Ports, corev1.ContainerPort{
+					Name:          "metrics",
+					ContainerPort: MetricsPort,
+					Protocol:      corev1.ProtocolTCP,
+				})
+			}
 		}
 
 		container.StartupProbe = &corev1.Probe{
@@ -565,6 +572,10 @@ func (feast *FeastServices) getContainerCommand(feastType FeastServiceType) []st
 	}
 
 	deploySettings := FeastServiceConstants[feastType]
+	deploySettings.Args = append([]string{}, deploySettings.Args...)
+	if feastType == OnlineFeastType && feast.isMetricsEnabled(feastType) {
+		deploySettings.Args = append([]string{deploySettings.Args[0], "--metrics"}, deploySettings.Args[1:]...)
+	}
 	targetPort := deploySettings.TargetHttpPort
 	tls := feast.getTlsConfigs(feastType)
 
@@ -732,6 +743,15 @@ func (feast *FeastServices) setService(svc *corev1.Service, feastType FeastServi
 		},
 	}
 
+	if feastType == OnlineFeastType && feast.isMetricsEnabled(feastType) {
+		svc.Spec.Ports = append(svc.Spec.Ports, corev1.ServicePort{
+			Name:       "metrics",
+			Port:       MetricsPort,
+			Protocol:   corev1.ProtocolTCP,
+			TargetPort: intstr.FromInt(int(MetricsPort)),
+		})
+	}
+
 	return controllerutil.SetControllerReference(feast.Handler.FeatureStore, svc, feast.Handler.Scheme)
 }
 
@@ -800,6 +820,18 @@ func (feast *FeastServices) getLogLevelForType(feastType FeastServiceType) *stri
 	return nil
 }
 
+func (feast *FeastServices) isMetricsEnabled(feastType FeastServiceType) bool {
+	if feastType != OnlineFeastType {
+		return false
+	}
+
+	if serviceConfigs := feast.getServerConfigs(feastType); serviceConfigs != nil && serviceConfigs.Metrics != nil {
+		return *serviceConfigs.Metrics
+	}
+
+	return false
+}
+
 func (feast *FeastServices) getNodeSelectorForType(feastType FeastServiceType) *map[string]string {
 	if serviceConfigs := feast.getServerConfigs(feastType); serviceConfigs != nil {
 		return serviceConfigs.ContainerConfigs.OptionalCtrConfigs.NodeSelector
@@ -1100,7 +1132,9 @@ func (feast *FeastServices) initRoute(feastType FeastServiceType) *routev1.Route
 }
 
 func applyCtrConfigs(container *corev1.Container, containerConfigs feastdevv1alpha1.ContainerConfigs) {
-	container.Image = *containerConfigs.DefaultCtrConfigs.Image
+	if containerConfigs.DefaultCtrConfigs.Image != nil {
+		container.Image = *containerConfigs.DefaultCtrConfigs.Image
+	}
 	// apply optional container configs
 	if containerConfigs.OptionalCtrConfigs.Env != nil {
 		container.Env = envOverride(container.Env, *containerConfigs.OptionalCtrConfigs.Env)

infra/feast-operator/internal/controller/services/services_test.go

@@ -23,6 +23,7 @@ import (
 	"github.com/feast-dev/feast/infra/feast-operator/internal/controller/handler"
 	. "github.com/onsi/ginkgo/v2"
 	. "github.com/onsi/gomega"
+	corev1 "k8s.io/api/core/v1"
 	metav1 "k8s.io/apimachinery/pkg/apis/meta/v1"
 	"k8s.io/apimachinery/pkg/types"
 	"k8s.io/apimachinery/pkg/util/intstr"
@@ -374,6 +375,49 @@ var _ = Describe("Registry Service", func() {
 			Expect(deployment.Spec.Template.Spec.NodeSelector).To(Equal(expectedNodeSelector))
 		})
 
+		It("should enable metrics on the online service when configured", func() {
+			featureStore.Spec.Services.OnlineStore = &feastdevv1alpha1.OnlineStore{
+				Server: &feastdevv1alpha1.ServerConfigs{Metrics: ptr(true)},
+			}
+
+			Expect(k8sClient.Update(ctx, featureStore)).To(Succeed())
+			Expect(feast.ApplyDefaults()).To(Succeed())
+			applySpecToStatus(featureStore)
+			feast.refreshFeatureStore(ctx, typeNamespacedName)
+
+			Expect(feast.deployFeastServiceByType(OnlineFeastType)).To(Succeed())
+
+			deployment := feast.initFeastDeploy()
+			Expect(deployment).NotTo(BeNil())
+			Expect(feast.setDeployment(deployment)).To(Succeed())
+
+			onlineContainer := GetOnlineContainer(*deployment)
+			Expect(onlineContainer).NotTo(BeNil())
+			Expect(onlineContainer.Command).To(Equal([]string{"feast", "serve", "--metrics", "-h", "0.0.0.0", "-p", "6566"}))
+			Expect(onlineContainer.Ports).To(ContainElement(corev1.ContainerPort{
+				Name:          "metrics",
+				ContainerPort: MetricsPort,
+				Protocol:      corev1.ProtocolTCP,
+			}))
+			metricsPortCount := 0
+			for _, port := range onlineContainer.Ports {
+				if port.Name == "metrics" {
+					metricsPortCount++
+				}
+			}
+			Expect(metricsPortCount).To(Equal(1))
+
+			svc := feast.initFeastSvc(OnlineFeastType)
+			Expect(svc).NotTo(BeNil())
+			Expect(feast.setService(svc, OnlineFeastType, false)).To(Succeed())
+			Expect(svc.Spec.Ports).To(ContainElement(corev1.ServicePort{
+				Name:       "metrics",
+				Port:       MetricsPort,
+				Protocol:   corev1.ProtocolTCP,
+				TargetPort: intstr.FromInt(int(MetricsPort)),
+			}))
+		})
+
 		It("should handle empty NodeSelector gracefully", func() {
 			// Set empty NodeSelector
 			emptyNodeSelector := map[string]string{}

infra/feast-operator/internal/controller/services/services_types.go

@@ -51,9 +51,10 @@ const (
 	caBundleAnnotation = "config.openshift.io/inject-trusted-cabundle"
 	caBundleName       = "odh-trusted-ca-bundle"
 
-	DefaultOfflineStorageRequest  = "20Gi"
-	DefaultOnlineStorageRequest   = "5Gi"
-	DefaultRegistryStorageRequest = "5Gi"
+	DefaultOfflineStorageRequest        = "20Gi"
+	DefaultOnlineStorageRequest         = "5Gi"
+	DefaultRegistryStorageRequest       = "5Gi"
+	MetricsPort                   int32 = 8000
 
 	AuthzFeastType    FeastServiceType = "authorization"
 	OfflineFeastType  FeastServiceType = "offline"