SikamikanikoBG/homelab-monitor v0.4.0

v0.4.0 — Prometheus /metrics endpoint

on GitHub

HomeLab Monitor can now feed your existing Prometheus + Grafana stack — while staying the same single, plug-and-play container.

✨ What's new

• /metrics Prometheus endpoint (port 9800). It's a pure translation layer over the in-memory snapshots the collector already keeps fresh — no extra polling, no double-sampling, and it degrades gracefully if prometheus_client isn't installed.
• 11 metrics exported: GPU VRAM (used/total), util, temp, power; host CPU / memory / per-mount disk; per-container state; per-systemd-unit state; and per-model loaded VRAM.
• Ready-to-import Grafana dashboard at docs/grafana/homelab_prometheus_dashboard.json (GPU, host, disk, and model VRAM in one view).
• README: new Prometheus integration section with a sample scrape config and import steps.

🙌 First contributor

Huge thanks to @turfin-logic for #5 — our very first community contribution, closing #3. Clean, well-scoped, and a genuine drop-in. 🎉

🔎 Try it

docker compose up -d --build
curl http://<your-host-ip>:9800/metrics

📋 Follow-up

• #6 — clear stale Prometheus label series on each scrape (good first issue).

Full changelog: v0.3.1...v0.4.0