WatchDog: n8n watches containers and restarts them

When many demos and services share one host, failures rarely scream — they go quiet. WatchDog is my n8n-based uptime guardian.

Problem

Manually checking whether KartonPak / AutoDocs / Technikalia are alive does not scale. You often learn about downtime only after someone messages you.

Architecture

1. Cron / interval in n8n (e.g. every 30–60s).
2. HTTP healthcheck for each service.
3. IF down → Docker container restart.
4. Alert to Discord (WatchDog bot).
5. POST status to the portfolio webhook: /api/watchdog.

Outcome

• Faster recovery from failed services
• Less manual server babysitting
• On wojciechdekowski.pl a live status feed proves the automation — not just a description

Lesson

Good monitoring is not a vanity dashboard. It is a closed loop: detect → repair → notify → show state.