Monitoring

Observability runs on the kube-prometheus-stack operator—a single HelmRelease that deploys Prometheus, Grafana, Alertmanager, and a full set of Kubernetes recording rules and dashboards.

Components

Prometheus scrapes metrics from all cluster components via ServiceMonitor and PodMonitor CRDs. Adding a new scrape target is as simple as creating a ServiceMonitor in the right namespace.

Grafana provides dashboards for cluster health, workload performance, and custom app metrics. Dashboards can be provisioned automatically via ConfigMaps labeled grafana_dashboard: "1". The Grafana sidecar picks them up without a pod restart.

Alertmanager handles alert routing. No external notification target is configured yet— mostly used to understand alert evaluation at this point.

Security Monitoring Metrics

Falco Talon exposes Prometheus metrics for every action it fires:

Metric	What it tracks
`falcosecurity_falco_talon_matches_total`	Falco events matched by Talon rules
`falcosecurity_falco_talon_actions_total`	Actions fired (with `status` label for success/fail)
`falcosecurity_falco_talon_notifications_total`	Notifications sent

These are scraped automatically via a ServiceMonitor on the Falco Talon service, and a bundled Grafana dashboard surfaces them under Dashboards → falco-talon.

Adding Custom Dashboards

Drop a ConfigMap into the monitoring namespace with the grafana_dashboard: "1" label and the JSON dashboard in a data key:

apiVersion: v1
kind: ConfigMap
metadata:
  name: my-dashboard
  namespace: monitoring
  labels:
    grafana_dashboard: "1"
data:
  my-dashboard.json: |
    { ... }

Grafana’s sidecar will detect and load it automatically.

Operators Databases