Observability Stack — Grafana LGTM + GlitchTip

New Issue

Self-hosted observability for the Familienarchiv application stack running on local hardware.

Goal: A separate docker-compose.observability.yml that can be started/stopped independently from the application stack, providing:

  • CPU/RAM dashboards per host and per container (Prometheus + Node Exporter + cAdvisor + Grafana)
  • Centralised log aggregation (Loki + Promtail)
  • Distributed request tracing (Tempo + OpenTelemetry)
  • Sentry-compatible error tracking for frontend and backend (GlitchTip)

Stack: Prometheus · Node Exporter · cAdvisor · Loki · Promtail · Tempo · Grafana · GlitchTip · Redis

Delivery order (suggested):

  1. Compose scaffold
  2. Metrics (Prometheus + Node Exporter + cAdvisor)
  3. Logs (Loki + Promtail)
  4. Traces (Tempo)
  5. Backend instrumentation (Micrometer + OTLP)
  6. Grafana (data sources + dashboards)
  7. GlitchTip infrastructure
  8. Frontend error tracking
  9. Backend error tracking
  10. Documentation
No due date
10% Completed