Complete Grafana dashboard and ship with Helm chart

[vriesdemichael]

Summary

The Helm chart has grafanaDashboard.enabled: false and the dashboard template appears incomplete. Users need visibility into operator behavior but no ready-to-use dashboard is provided.

Problem

Current state:

• values.yaml has grafanaDashboard.enabled: false
• Dashboard ConfigMap template exists but content is minimal
• Documentation references monitoring but no dashboard JSON is shipped
• Users must create their own dashboards from scratch

Proposed Solution

1. Create Comprehensive Dashboard JSON

Dashboard should include panels for:

Reconciliation Overview

• Reconciliation rate (success/failure)
• Reconciliation duration (p50, p95, p99)
• Active resources by type and phase
• Error rate by error type

Rate Limiting

• Tokens available (global and per-namespace)
• Wait time histogram
• Timeout count

Drift Detection

• Orphaned resources count
• Unmanaged resources count
• Last successful drift check

Health

• Leader election status
• Database connection status
• Keycloak instance status

2. Add to Helm Chart

# templates/grafana-dashboard-configmap.yaml
{{- if .Values.monitoring.grafanaDashboard.enabled }}
apiVersion: v1
kind: ConfigMap
metadata:
  name: {{ include "keycloak-operator.fullname" . }}-dashboard
  labels:
    {{- include "keycloak-operator.labels" . | nindent 4 }}
    grafana_dashboard: "1"
data:
  keycloak-operator.json: |
    {{ .Files.Get "dashboards/keycloak-operator.json" | indent 4 }}
{{- end }}

3. Enable by Default

Change default in values.yaml:

monitoring:
  grafanaDashboard:
    enabled: true  # Changed from false

Acceptance Criteria

• Dashboard JSON created with all key metrics
• Dashboard tested with Grafana 9.x and 10.x
• Configmap template properly includes dashboard
• Enabled by default
• Screenshot added to documentation

References

• ADR-048: Prometheus metrics exposure
• Review finding: TODO/SENIOR_SRE_REVIEW.md