Monitoring Homelab

A full monitoring stack on a 3-node Proxmox VE cluster running on consumer hardware. Every config file written by hand so I actually understand what each setting does.

What’s running

• Prometheus scraping 15s metrics from all nodes and services
• Grafana dashboards (node health, networking, container)
• Alertmanager routing to PagerDuty
• node_exporter on every host
• blackbox_exporter for HTTP/TCP endpoint checks
• Tailscale mesh for remote access

Alert rules

Seven alert rules in production:

• NodeDown — any host unreachable for > 2 min
• DiskSpaceLow — < 15% free
• HighCPU — > 90% sustained 10 min
• HighRAM — > 90% sustained 5 min
• EndpointDown — any blackbox target failing
• ServiceCrashed — systemd unit in failed state
• ProxmoxStorageWarn — storage pool > 85%

Paging

Alertmanager → PagerDuty. SMS on alert, voice escalation if not acknowledged within 5 minutes. Tested by triggering each rule deliberately.

Repo

github.com/E-mma9/Monitoring-Homelab