ocp-grafana-operator

name	ocp-grafana-operator
metadata	{"author":"rhoai3-demo","version":"1.1.0","platform-family":"ocp","platform-baseline":"repo","ocp-baseline":"repo","skill-group":"OpenShift Platform"}
description	Use when documenting, reviewing, or rebuilding GitOps-managed Grafana on OpenShift for the rhoai3-demo: Grafana Operator Subscription from the community operator catalog, Grafana custom resources, OpenShift OAuth proxy route posture, service serving certificates, injected CA bundles, session secrets, GrafanaDatasource resources, GrafanaDashboard resources, cluster-monitoring-view access to Thanos Querier, Argo CD sync options, namespace and OperatorGroup handling, official Grafana Operator docs, API reference checks, Kustomize installation boundaries, Argo CD dashboard GitOps patterns, and Red Hat CoP grafana-operator catalog curation. Do NOT use for RHOAI model monitoring semantics, vLLM/KServe metrics, or dashboard content design; use rhoai-model-management-monitoring. Do NOT treat the CoP catalog as product authority; verify fields with official Grafana Operator docs, official OCP docs, installed OLM package metadata, and active CRDs.

OCP Grafana Operator

Use this skill to deploy and manage Grafana through GitOps for the rhoai3-demo observability layer. It captures the Red Hat Community of Practice grafana-operator catalog pattern and adapts it to the repo's local-curation and OpenShift safety model.

Source Grounding

Read references/official-grafana-operator-docs.md before validating Grafana Operator CRs, datasource CRs, dashboard CRs, or Argo CD dashboard GitOps behavior. Use it to distinguish official Grafana Operator behavior from repo-specific OpenShift/OLM adaptation.

Read references/source-capture.md before using product behavior. The CoP catalog is a GitOps pattern source, not product support authority. The Grafana Operator in the captured catalog is installed from community-operators, so channel, install mode, custom resource fields, images, and support posture must be verified from the active cluster catalog, CRDs, official Grafana Operator docs, and relevant OpenShift docs.

Read references/gitops-catalog-grafana-pattern.md when rebuilding operator, instance, OAuth proxy, datasource, dashboard, or aggregate overlays.

Demo Grafana Posture

For this AWS-hosted RHOAI demo:

• Treat Grafana as an optional demo observability UI, introduced only when a step needs dashboards for platform, model-serving, GPU, or application metrics.
• Curate the CoP pattern locally. Do not commit remote Kustomize references to redhat-cop/gitops-catalog.
• Prefer a dedicated namespace for the Grafana Operator and instance. Verify there is only one OperatorGroup in that namespace.
• Keep operator install resources separate from Grafana instance, datasource, dashboard, route, RBAC, and secret resources unless a single aggregate Argo CD Application is deliberately selected.
• Use OpenShift OAuth and route TLS for browser access. Do not copy anonymous or basic-auth settings without documenting the demo-specific reason.
• Never commit generated service-account tokens, real session secrets, API keys, or datasource credentials.
• Use cluster-monitoring-view or narrower RBAC only when the Grafana service account needs access to OpenShift monitoring data.
• Keep dashboard content ownership in the component skill that owns the metric semantics. For model-serving dashboards, use rhoai-model-management-monitoring.

Grafana GitOps Model

Use the CoP pattern to frame:

• Operator layer: OLM Subscription for package grafana-operator from community-operators.
• Namespace and OperatorGroup layer: namespace-specific install scope supplied by the deployable overlay.
• Instance layer: Grafana custom resource plus service account, OAuth proxy, serving certificate, injected CA bundle, session secret, route, and RBAC resources.
• Datasource layer: GrafanaDatasource for OpenShift monitoring or user workload monitoring endpoints, with token handling reviewed as sensitive.
• Dashboard layer: GrafanaDashboard or operator-supported dashboard resources after CRDs and selectors are verified.
• Aggregate layer: optional overlay that combines operator and instance resources with Argo CD dry-run handling while CRDs become available.

Workflow

1. Confirm the active OpenShift baseline in docs/PLATFORM_BASELINE.md.
2. Read references/official-grafana-operator-docs.md.
3. Read references/source-capture.md.
4. Read references/gitops-catalog-grafana-pattern.md.
5. Identify whether the task concerns:
   • Grafana Operator installation or lifecycle
   • namespace and OperatorGroup scope
   • Grafana instance configuration
   • OAuth proxy, Route, service serving certificate, or injected CA bundle
   • datasource access to OpenShift monitoring or user workload monitoring
   • dashboard CRs and dashboard selectors
   • RBAC, service-account tokens, session secrets, or external credentials
6. Pair with project-red-hat-operator-gitops for layout and lifecycle management, and with project-gitops-authoring for repo-specific Argo CD Application standards.
7. Verify all API versions, CR fields, channel values, images, route settings, RBAC, datasource fields, dashboard fields, and secret handling against the active cluster schema and official Grafana Operator API reference before committing manifests.
8. For live operations, use the repo environment guard and pair this skill with env-troubleshoot, env-manage-resources, or env-deploy-and-evaluate.
9. Validate the output with references/validation-checklist.md.

Related Skills

• Use project-red-hat-operator-gitops for the local operator, instance, and aggregate GitOps pattern.
• Use project-gitops-authoring for repo-specific Kustomize and Argo CD Application conventions.
• Use project-manifest-review for structural and security review of Kubernetes, OpenShift, and Argo CD manifests.
• Use ocp-observability for OCP monitoring and user workload monitoring concepts.
• Use ocp-ingress-gateway-routes for Route and TLS behavior.
• Use ocp-security-rbac-scc for service accounts, ClusterRoles, ClusterRoleBindings, RoleBindings, and token review permissions.
• Use rhoai-model-management-monitoring for RHOAI model-serving metrics and dashboard content.

References

• references/source-capture.md
• references/official-grafana-operator-docs.md
• references/gitops-catalog-grafana-pattern.md
• references/validation-checklist.md
• examples/grafana-review-patterns.md