Heartbeat

Periodic check-in tasks for sepia. Run these on a schedule (daily/weekly) to catch issues early.

Daily

• Check for unhealthy containers: docker ps --filter "health=unhealthy"
• Review borgmatic logs: docker logs borgmatic | tail -20
• Check disk space: df -h / /media
• Verify core services are up: homeassistant, seafile, caddy

Weekly

• Check backup repository sizes: du -sh /opt/borgmatic/cache
• Review Grafana for anomalies (CPU, memory, network trends)
• Check for Docker image updates (critical services: Caddy, Home Assistant, PostgreSQL)
• Verify SSL certificate expiry (Caddy auto-renews, but confirm): docker exec caddy caddy list-modules | grep tls

Monthly

• Check for obsolete containers/images: docker system df -v
• Audit open ports and Caddy routes for services that no longer exist
• Review and rotate old journal entries in journal/
• Check DSMR reader backups in /opt/dsmr/backups/

On-Demand

• After any compose change: verify all services healthy
• After host reboot: check all expected containers started
• After backup failures: run manual backup and investigate

Cleanup Reminders

• 2026-05-31 — Prune Hermes backup: /media/scratch/hermes-pre-cleanup-2026-05-17.tar.gz