Grafana Dashboard Improvements Design

Date: 2026-01-11 Issue: selfhosted-cluster-78q Status: Approved

Goals

1. Operational visibility - At-a-glance health of all services
2. Troubleshooting support - Deep-dive dashboards for debugging
3. Consistency - Unified provisioning method and organization

Decisions

Decision	Choice
Folder organization	By layer (Infrastructure, Platform, Applications, Kubernetes, Logs)
Dashboard sources	Grafana.com with pinned revisions; GitHub for project-maintained dashboards
Provisioning method	GrafanaDashboard CRD (migrate Velero from ConfigMap)
Directory rename	system-dashboards/ → service-ingresses/

New Dashboards

Service	Source	Dashboard ID/URL	Folder
Alloy	GitHub	grafana/alloy repo	Infrastructure
Longhorn	Grafana.com	ID 16888 rev 9	Infrastructure
Traefik	Grafana.com	ID 17346 rev 10	Infrastructure
CNPG/PostgreSQL	GitHub	cloudnative-pg/grafana-dashboards	Platform
Vault	Grafana.com	ID 12904 rev 2	Platform
MetalLB	Grafana.com	ID 14127 rev 1	Infrastructure

Migrations

Current Dashboard	Action
Velero (ConfigMap)	Replace with Grafana.com ID 11055 rev 2 → Applications
ArgoCD	Move to Applications folder
NATS	Move to Platform folder
Kubernetes Overview	Move to Kubernetes folder (rename to api-server)
Loki Logs	Move to Logs folder
Node Exporter	Move from monitoring-prometheus/ to Infrastructure
Cert-Manager	Move from cert-manager-config/ to Platform
Temporal (Terraform)	Update folder to “Applications”

Repository Structure

argocd/app-configs/grafana/dashboards/
├── kustomization.yaml
├── infrastructure/
│   ├── alloy.yaml
│   ├── longhorn.yaml
│   ├── traefik.yaml
│   ├── metallb.yaml
│   └── node-exporter.yaml
├── platform/
│   ├── cnpg.yaml
│   ├── nats.yaml
│   ├── vault.yaml
│   └── cert-manager.yaml
├── applications/
│   ├── argocd.yaml
│   └── velero.yaml
├── kubernetes/
│   └── api-server.yaml
└── logs/
    └── loki-overview.yaml

CRD Template

apiVersion: grafana.integreatly.org/v1beta1
kind: GrafanaDashboard
metadata:
  name: <service-name>
  namespace: grafana
  labels:
    grafana.integreatly.org/instance: grafana
spec:
  instanceSelector:
    matchLabels:
      dashboards: grafana
  folder: <Infrastructure|Platform|Applications|Kubernetes|Logs>
  url: https://grafana.com/api/dashboards/<ID>/revisions/<REV>/download

For GitHub-sourced dashboards:

  url: https://raw.githubusercontent.com/<org>/<repo>/<ref>/<path>.json

Cleanup Actions

1. Delete argocd/app-configs/velero/monitoring/grafana-dashboard.yaml
2. Rename argocd/app-configs/system-dashboards/ → argocd/app-configs/service-ingresses/
3. Update references in cluster-app if needed

Out of Scope

• Authentik dashboard (limited metrics available)
• Additional alert rules (separate effort)