Grafana¶

Example Grafana dashboards, provided as is, are hosted on grafana.com. Container orchestrator specific instructions to enable these dashboards are available below.

Red Hat Advanced Cluster Management¶

Red Hat Advanced Cluster Management (ACM) includes an Observability component that provides multi-cluster monitoring with a built-in Grafana instance. Custom dashboards can be loaded into this Grafana instance via Kubernetes ConfigMap resources. The dashboards hosted on grafana.com require several modifications to work correctly in this environment.

Dashboard ConfigMap¶

ACM Grafana discovers custom dashboards from ConfigMap resources in the open-cluster-management-observability namespace that carry the grafana-custom-dashboard: "true" label. The dashboard JSON is stored as a data key ending in .json.

---
apiVersion: v1
kind: ConfigMap
metadata:
  name: hpe-array-exporter-dashboard
  namespace: open-cluster-management-observability
  labels:
    cluster.open-cluster-management.io/backup: ""
    grafana-custom-dashboard: "true"
data:
  hpe-exporter-dashboard.json: |
    {
      # Dashboard JSON content goes here
    }

Adapting Grafana.com Dashboards¶

Dashboards exported from grafana.com include sections and settings that are not compatible with ACM Grafana. The following modifications are required when wrapping a grafana.com dashboard export into a ConfigMap.

Remove export-only sections. The __inputs and __requires sections at the top of the dashboard JSON are used during manual Grafana UI imports and serve no purpose in a ConfigMap-based deployment. These sections should be removed entirely.

Fix datasource references. Grafana.com exports reference a datasource variable such as ${DS_PROMETHEUS} that relies on the __inputs substitution mechanism. This substitution does not occur when the dashboard is loaded via a ConfigMap. All datasource references must be replaced with an explicit object reference.

Replace any occurrence of "datasource": null or "datasource": "${DS_PROMETHEUS}" with:

"datasource": {"type": "prometheus", "uid": "default"}

Update annotation datasource. The annotation datasource "datasource": "-- Grafana --" is a deprecated string format. Replace it with the object format:

"datasource": {"type": "datasource", "uid": "grafana"}

Update schema version. Grafana.com exports may use an older schemaVersion. Update the value to match the version supported by the ACM Grafana instance (e.g. 39).

Remove stale fields. Fields such as gnetId, iteration, and "style": "dark" are grafana.com export artifacts and can be removed.

Metrics Allowlist¶

ACM Observability collects metrics from managed clusters and forwards them to the hub cluster. Only metrics included in a custom allowlist ConfigMap are collected. The exporter runs as a user workload, so its metrics must be added to the uwl_metrics_list.yaml section of the allowlist rather than the metrics_list.yaml section.

The following example adds a match rule for HPE Alletra Storage MP B10000 metrics to the allowlist ConfigMap on the hub cluster.

---
kind: ConfigMap
apiVersion: v1
metadata:
  name: observability-metrics-custom-allowlist
  namespace: open-cluster-management-observability
data:
  uwl_metrics_list.yaml: |
    matches:
    - __name__=~"hpealletrastoragemp_.*"