Platform Monitoring
Managed cluster dashboards
Native managed clusters with the current observability add-on include five read-only dashboards in Grafana’s Cluster folder. Each shows the registered cluster ID so that open tabs remain distinguishable.
| Dashboard | Use it to inspect |
|---|---|
| Cluster overview | Node readiness, capacity, workload availability and collection health |
| Nodes and capacity | CPU, memory, requests, limits and node pressure |
| Workloads and pod health | Availability, pending pods, restarts and resource use by namespace |
| Cluster System and add-ons | Observed Kubernetes readiness and reconciliation of installed components |
| Observability pipeline | Scrapes, Prometheus, alert evaluation, logs and Grafana |
Start with Inspect cluster health. The Portal Grafana route works before cluster ingress is configured. Every member of the selected Context can view its cluster telemetry; dashboard filters do not grant or restrict access.
These dashboards use cluster-local metrics and the existing Loki data source. They add no VPA dependency, alert rules, notification channels or synthetic uptime probes. A component being ready does not prove end-to-end authentication, report delivery, VPN connectivity or successful image transfers.
Legacy platform monitoring
The predecessor Core Platform Infra platform-monitoring feature has a separate
monitoring installation. Its Global, Namespace, Continuous Load and Platform
Liveness dashboards depend on that feature’s collectors and probes. They are not
installed by the managed observability add-on and should not be assumed available
on a native managed cluster.
For legacy application monitoring and shared monitoring installations, see Application monitoring.