A practical home server monitoring setup with alerts

A home server setup uses on each machine to collect basic health data, while Prometheus, Alertmanager, and Blackbox Exporter run together inside one LXC. Grafana turns the collected data into a visible dashboard, with a small office screen planned for always-on monitoring. Alerts are sent to PagerDuty.

There are 7 alert rules covering disk use, memory use, CPU use, a machine going offline, and checks that real services still respond. The checked services include , Pi-hole, NPM, and Paperless. The actual s are available in a .

A possible next step is an external heartbeat check aimed at a outside the cluster, so failures in the monitoring setup itself can also be noticed.

Key points

  • runs on each host to collect machine health data.
  • Prometheus, Alertmanager, and Blackbox Exporter run inside one LXC.
  • Grafana is used to build a dashboard that could stay visible on a small screen.
  • Alerts cover disk, memory, CPU, offline machines, and failed service checks.
  • An external heartbeat check with a outside the cluster is being considered.
Read original