feat: add new metrics required by release dashboard and SLO (#4437)

this commit adds new metrics to enable a new SLO and Dashboard for the Release Service Signed-off-by: Leandro Mendes <[email protected]>
redhat-appstudio · Sep 3, 2024 · 217dc75 · 217dc75
1 parent 81b9fa6
commit 217dc75
Showing 1 changed file with 11 additions and 3 deletions.
diff --git a/components/monitoring/prometheus/base/monitoringstack/monitoringstack.yaml b/components/monitoring/prometheus/base/monitoringstack/monitoringstack.yaml
@@ -105,6 +105,7 @@ spec:
         - '{__name__="tekton_pipelines_controller_running_taskruns_throttled_by_node"}'
         - '{__name__="tekton_pipelines_controller_pipelinerun_duration_seconds_sum"}'
         - '{__name__="tekton_pipelines_controller_pipelinerun_duration_seconds_count"}'
+        - '{__name__="tekton_pipelines_controller_pipelinerun_taskrun_duration_seconds_bucket"}'
         - '{__name__="watcher_workqueue_depth"}'
         - '{__name__="watcher_client_latency_bucket"}'
         - '{__name__="pac_watcher_work_queue_depth"}'
@@ -114,9 +115,12 @@ spec:
         - '{__name__="controller_runtime_reconcile_errors_total", namespace!~".*-tenant|openshift-.*|kube-.*"}'
         - '{__name__="controller_runtime_reconcile_total", namespace!~".*-tenant|openshift-.*|kube-.*"}'
         - '{__name__="kube_pod_status_unschedulable", namespace!~".*-tenant|openshift-.*|kube-.*"}'
-        - '{__name__="kube_pod_container_status_restarts_total", namespace="openshift-pipelines"}'
-        - '{__name__="kube_pod_container_status_waiting_reason", namespace!~".*-tenant|openshift-.*|kube-.*"}'
-        - '{__name__="kube_pod_status_phase", namespace!~".*-tenant|openshift-.*|kube-.*"}'
+        - '{__name__="kube_pod_container_status_restarts_total", namespace="openshift-pipelines|release-service"}'
+        - '{__name__="kube_pod_container_status_waiting_reason", namespace!~".*-tenant|openshift-.*|kube-.*|release-service"}'
+        - '{__name__="kube_pod_status_phase", namespace!~".*-tenant|openshift-.*|kube-.*|release-service"}'
+        - '{__name__="kube_pod_container_status_terminated_reason", namespace="release-service"}'
+        - '{__name__="kube_pod_container_status_last_terminated_reason", namespace="release-service"}'
+        - '{__name__="kube_pod_container_status_ready", namespace="release-service"}'
         - '{__name__="kube_persistentvolume_status_phase", namespace!~".*-tenant|openshift-.*|kube-.*"}'
         - '{__name__="kube_resourcequota", namespace!~".*-tenant|openshift-.*|kube-.*"}'
         - '{__name__="kube_statefulset_status_replicas_ready", namespace="gitops-service-argocd"}'
@@ -127,6 +131,10 @@ spec:
         - '{__name__="kube_deployment_status_replicas_available", namespace=~"smee.*"}'
         - '{__name__="argocd_app_reconcile_bucket", namespace="gitops-service-argocd"}'
         - '{__name__="argocd_app_info", namespace="gitops-service-argocd"}'
+        - '{__name__="container_cpu_usage_seconds_total", namespace="release-service"}'
+        - '{__name__="container_memory_usage_bytes", namespace="release-service"}'
+        - '{__name__="kube_pod_container_resource_limits", namespace="release-service"}'
+
     relabelings:
     # override the target's address by the prometheus-k8s service name.
     - action: replace