See problems before they hurt
What you don’t measure, you can’t operate. Metrics, logs and synthetic checks flow into one place, dashboards show the whole health picture at a glance, and alerts speak up before anyone notices.
Central metrics
Prometheus gathers time series from the host, containers, databases and the NAS into one central TSDB — fed by several exporters.
Logs in one place
An agent collects the logs of every container and files them in a shared store. They are searchable in the same dashboard as the metrics — no jumping between tools when a curve needs explaining.
Dashboards
Grafana visualizes everything, secured by the same single sign-on. One glance covers CPU, memory, temperature and service health.
Synthetic checks
Blackbox probes test every service for HTTP reachability and TLS validity; Uptime Kuma adds an external status page.
Alerts
Thresholds trigger push notifications — a full disk, a dead service or an expiring certificate report themselves.
Self-healing after a reboot
After a reboot, the network drives wait until the NAS really answers. Then, and every hour, a watchdog checks whether a container still holds an empty, stale view of its data, and restarts exactly that one. If a public entry is switched off in the reverse proxy, an alert goes out. Every watchdog reports its own run via a dead man’s switch — the self-healing is monitored too.