Change ServiceMonitor default labels

Author: Nitish Bhat

bundle/manifests/amd-gpu-operator.clusterserviceversion.yaml

@@ -32,7 +32,7 @@ metadata:
     capabilities: Seamless Upgrades
     categories: AI/Machine Learning,Monitoring
     containerImage: docker.io/rocm/gpu-operator:v1.2.0
-    createdAt: "2025-05-04T02:40:22Z"
+    createdAt: "2025-05-08T06:33:46Z"
     description: |-
       Operator responsible for deploying AMD GPU kernel drivers, device plugin, device test runner and device metrics exporter
       For more information, visit [documentation](https://instinct.docs.amd.com/projects/gpu-operator/en/latest/)
@@ -461,7 +461,7 @@ spec:
         x-descriptors:
         - urn:alm:descriptor:com.amd.deviceconfigs:enable
       - description: HonorLabels chooses the metric's labels on collisions with target
-          labels (default false)
+          labels (default true)
         displayName: HonorLabels
         path: metricsExporter.prometheus.serviceMonitor.honorLabels
         x-descriptors:

bundle/manifests/amd.com_deviceconfigs.yaml

@@ -653,8 +653,9 @@ spec:
                               (default false)
                             type: boolean
                           honorLabels:
+                            default: true
                             description: HonorLabels chooses the metric's labels on
-                              collisions with target labels (default false)
+                              collisions with target labels (default true)
                             type: boolean
                           honorTimestamps:
                             description: HonorTimestamps controls whether the scrape

config/crd/bases/amd.com_deviceconfigs.yaml

@@ -649,8 +649,9 @@ spec:
                               (default false)
                             type: boolean
                           honorLabels:
+                            default: true
                             description: HonorLabels chooses the metric's labels on
-                              collisions with target labels (default false)
+                              collisions with target labels (default true)
                             type: boolean
                           honorTimestamps:
                             description: HonorTimestamps controls whether the scrape

config/manifests/bases/amd-gpu-operator.clusterserviceversion.yaml

@@ -432,7 +432,7 @@ spec:
         x-descriptors:
         - urn:alm:descriptor:com.amd.deviceconfigs:enable
       - description: HonorLabels chooses the metric's labels on collisions with target
-          labels (default false)
+          labels (default true)
         displayName: HonorLabels
         path: metricsExporter.prometheus.serviceMonitor.honorLabels
         x-descriptors:

docs/sphinx/_toc.yml

@@ -36,6 +36,7 @@ subtrees:
     entries:
       - file: metrics/exporter
       - file: metrics/kube-rbac-proxy
+      - file: metrics/prometheus
       - file: metrics/health
         title: Health Checks
   - caption: Test Runner

docs/sphinx/_toc.yml.in

@@ -36,6 +36,7 @@ subtrees:
     entries:
       - file: metrics/exporter
       - file: metrics/kube-rbac-proxy
+      - file: metrics/prometheus
       - file: metrics/health
         title: Health Checks
   - caption: Test Runner

helm-charts-k8s/Chart.lock

@@ -6,4 +6,4 @@ dependencies:
   repository: file://./charts/kmm
   version: v1.0.0
 digest: sha256:f9a315dd2ce3d515ebf28c8e9a6a82158b493ca2686439ec381487761261b597
-generated: "2025-05-04T02:28:56.201802096Z"
+generated: "2025-05-08T06:33:22.365680047Z"

helm-charts-k8s/README.md

@@ -66,7 +66,8 @@ helm repo update
 ```bash
 helm install amd-gpu-operator rocm/gpu-operator-charts \
   --namespace kube-amd-gpu \
-  --create-namespace
+  --create-namespace \
+  --version=v1.2.0
 ```
 
 #### Installation Options

helm-charts-k8s/crds/deviceconfig-crd.yaml

@@ -656,8 +656,9 @@ spec:
                               false)
                             type: boolean
                           honorLabels:
+                            default: true
                             description: HonorLabels chooses the metric's labels on
-                              collisions with target labels (default false)
+                              collisions with target labels (default true)
                             type: boolean
                           honorTimestamps:
                             description: HonorTimestamps controls whether the scrape

helm-charts-openshift/Chart.lock

@@ -6,4 +6,4 @@ dependencies:
   repository: file://./charts/kmm
   version: v1.0.0
 digest: sha256:25200c34a5cc846a1275e5bf3fc637b19e909dc68de938189c5278d77d03f5ac
-generated: "2025-04-30T04:46:12.083919622Z"
+generated: "2025-05-08T06:33:43.637209508Z"

helm-charts-openshift/crds/deviceconfig-crd.yaml

@@ -656,8 +656,9 @@ spec:
                               false)
                             type: boolean
                           honorLabels:
+                            default: true
                             description: HonorLabels chooses the metric's labels on
-                              collisions with target labels (default false)
+                              collisions with target labels (default true)
                             type: boolean
                           honorTimestamps:
                             description: HonorTimestamps controls whether the scrape

internal/metricsexporter/metricsexporter.go

@@ -56,14 +56,15 @@ const (
 	defaultInitContainerImage         = "busybox:1.36"
 	servicePort                 int32 = 5000
 	nobodyUser                        = 65532
-	ExporterName                      = "amd-device-metrics-exporter"
+	ExporterName                      = "metrics-exporter"
 	KubeRbacName                      = "kube-rbac-proxy"
 	StaticAuthSecretName              = ExporterName + "-static-auth-config"
 	defaultSAName                     = "amd-gpu-operator-metrics-exporter"
 	kubeRbacSAName                    = "amd-gpu-operator-metrics-exporter-rbac-proxy"
 	svcLabel                          = "app.kubernetes.io/service"
 )
 
+var serviceMonitorLabelPair = []string{"app", "amd-device-metrics-exporter"}
 var metricsExporterLabelPair = []string{"app.kubernetes.io/name", ExporterName}
 
 //go:generate mockgen -source=metricsexporter.go -package=metricsexporter -destination=mock_metricsexporter.go MetricsExporter
@@ -585,18 +586,16 @@ func (nl *metricsExporter) SetServiceMonitorAsDesired(sm *monitoringv1.ServiceMo
 		AttachMetadata:    devConfig.Spec.MetricsExporter.Prometheus.ServiceMonitor.AttachMetadata,
 	}
 
-	// Set custom labels or use default
-	if sm.Labels == nil {
-		sm.Labels = make(map[string]string)
+	// Set custom labels
+	sm.Labels = map[string]string{
+		serviceMonitorLabelPair[0]: serviceMonitorLabelPair[1],
 	}
+
 	if len(devConfig.Spec.MetricsExporter.Prometheus.ServiceMonitor.Labels) > 0 {
 		// Use custom labels from the CRD
 		for k, v := range devConfig.Spec.MetricsExporter.Prometheus.ServiceMonitor.Labels {
 			sm.Labels[k] = v
 		}
-	} else {
-		// Use default labels
-		sm.Labels["release"] = "prometheus-operator"
 	}
 
 	return controllerutil.SetControllerReference(devConfig, sm, nl.scheme)