monitoring-authoring

To create PrometheusRule alerts, you author manifests that apply the required kube-prometheus-stack discovery label and standardized alert annotations, including severity and for-duration guidance. This ensures Prometheus correctly discovers and executes your alerting and recording rules for Kubernetes workloads.

Configuring ServiceMonitor and PodMonitor for metric scraping requires setting correct namespace placement, selectors, and required discovery labels. This ensures Prometheus can properly discover and scrape metrics from your targeted Kubernetes services and pods.

The best way to route AlertmanagerConfig and create Silence CRs is to declaratively add routing configurations and suppression custom resources. This approach aligns suppression behavior with platform conventions for known alerts without breaking discovery labels.

Writing Canary health checks involves creating Canary resources for synthetic HTTP checks and Kubernetes CEL-based health verification. You must avoid canary patterns that produce false negatives to ensure accurate synthetic monitoring of application and platform components.

PrometheusRules or ServiceMonitors may not be discovered if they lack required kube-prometheus-stack discovery labels or use incorrect Kubernetes placement and selectors. Authoring monitoring resources must enforce these required labels and correct selectors for proper discovery.