Building a Production-Grade Observability Stack for My Homelab
I run a self-hosted media server at home, Jellyfin for movies and TV, Immich for photos, Paperless for documents, and a handful of other services. For a long time I had no real visibility into what any of it was doing. If something broke, I’d find out when it stopped working. So I decided to fix that with a real observability stack. The Goal Metrics: CPU, memory, disk, network, and GPU utilization over time Logs: aggregated from all containers in one place Alerts: that actually reach me when something goes wrong Auto-recovery: containers that restart themselves without my intervention The Stack Everything runs as Docker containers managed by a single docker-compose.yml. The architecture is: ...