HomeLab Monitor

What is HomeLab Monitor about?

HomeLab Monitor provides a unified, browser‑based view of an entire home‑lab or AI rig. It aggregates data from GPUs, Docker containers, systemd services, and filesystem usage on any number of machines reachable via SSH—Linux, Raspberry Pi, or Windows—without requiring agents on the remote hosts.

How to use HomeLab Monitor?

1. Deploy – Download the provided docker-compose.yml and start the stack:

   curl -fsSLO https://raw.githubusercontent.com/SikamikanikoBG/homelab-monitor/main/docker-compose.yml
   docker compose up -d
   

2. Access – Open http://<your‑host>:9800 in a browser.
3. Add hosts – In the Hosts tab, paste the auto‑generated SSH key onto each remote machine; the hub will poll them automatically.
4. Enable MCP (optional) – The built‑in read‑only MCP server runs on port 9810. Set ENABLE_MCP=1 (default) in the compose environment to expose it. AI agents can then connect using a URL like http://YOUR-HUB:9810/mcp.
5. Configure – Adjust optional environment variables such as SAMPLE_INTERVAL, RETENTION_DAYS, PRESSURE_FREE_MB, PORT, and alert settings via the docker-compose.yml environment: block.

Key features of HomeLab Monitor

• GPU insights – Live VRAM, utilization, power, temperature, and automatic attribution of which container is using the GPU.
• Container health – Real resident RAM and VRAM per container, OOM detection, and status indicators.
• Systemd service view – Show local and remote units, highlight failures, and allow custom watch lists.
• Disk treemaps – Interactive WizTree‑style visualisation to locate space hogs.
• Multi‑host support over SSH – No agents; a tiny probe is transferred on demand.
• Alerting – Edge‑triggered Discord and ntfy.sh notifications.
• Prometheus endpoint – /metrics for external scraping, plus a ready‑made Grafana dashboard.
• Built‑in read‑only MCP server – Exposes the same telemetry to AI agents via 12 named tools.
• Zero‑GPU requirement – The UI works even if no GPU is present; GPU panels simply stay inactive.

Use cases of HomeLab Monitor

• Home‑lab operations – Quickly identify which model server is occupying the GPU, which container is leaking RAM, or which disks are filling up.
• AI‑rig management – Monitor multiple LLM serving back‑ends (Ollama, vLLM, llama.cpp, A1111, etc.) and correlate usage with specific containers.
• Remote troubleshooting – Access a full health dashboard from a phone over VPN, even for headless Pi or Windows hosts.
• Automated AI assistance – Connect Claude or other MCP‑compatible agents to let them answer operational questions without manual UI navigation.
• Alert-driven workflows – Receive Discord/ntfy alerts when GPU pressure exceeds thresholds or services fail.

FAQ from the HomeLab Monitor

Q: Do I need a GPU on the host running the dashboard?
A: No. The dashboard works without a GPU; GPU panels become active only when a GPU is detected on any monitored host.

Q: How is data collected from remote machines?
A: The hub sends a lightweight Python/PowerShell probe over SSH; nothing is persisted on the remote side.

Q: Is the MCP server writable?
A: It is read‑only by design; there are no write tools, so agents can only query data.

Q: Can I disable the MCP server?
A: Yes, set ENABLE_MCP=0 in the compose environment to run only the dashboard.

Q: What storage does the dashboard use?
A: SQLite files (gpu.db, etc.) are stored under ./data and survive container restarts.

Q: How are alerts configured?
A: Through environment variables and the web UI; supported targets include Discord webhooks and ntfy.sh endpoints.

Q: Is there any cloud dependency?
A: No. All components run locally; the only outbound call is an optional daily GitHub‑releases check, which can be disabled via CHECK_UPDATES=false.

🛰️ HomeLab Monitor

One page for your whole home lab & AI rig — GPU, containers, services, disks. No agents, no Prometheus/Grafana, no cloud.

Your home lab grew into a couple of machines, a Pi, and a GPU that's mysteriously always busy. HomeLab Monitor gives you one self-hosted page that answers the real questions: which model is holding the GPU, which container is eating RAM, what's filling your disks, and is anything down — across every box over SSH: Linux, a Pi, even Windows. Readable from your phone over the VPN.

Get started

# Grab the compose file and go. No GPU required — the GPU panels just light up when one's present.
curl -fsSLO https://raw.githubusercontent.com/SikamikanikoBG/homelab-monitor/main/docker-compose.yml
docker compose up -d

Open http://<your-host>:9800 and you're done. Full options (from source, GPU toolkit, Windows/WSL2) → Install docs.

🆕 v0.14.0 — a built-in read-only MCP server: connect Claude (or any MCP client) to your homelab and explore it with full dashboard parity, no extra container. Release notes · changelog · MCP docs.

What you get

• GPU, demystified — live VRAM/util/power/temp, and which container is holding the card (auto-mapped).
• Containers, honestly — health plus RAM and VRAM in separate columns (real resident RAM, not page cache).
• systemd services — local or remote, your own units highlighted, failures first.
• WizTree-style disk treemaps — scan a filesystem, drill into folders, find the space hogs.
• Multi-machine over SSH — paste one key per box; Linux, a Pi, even Windows. No agents, no installs.
• Push alerts — Discord and ntfy.sh, edge-triggered so they don't spam.

Full tab-by-tab tour → Features.

Multi-machine, in two sentences

Open the Hosts tab, paste the hub's auto-generated SSH key onto each remote, and the hub starts polling it — no agents, just SSH + Python 3 (PowerShell on Windows). The hub pipes a small self-contained probe over SSH; nothing persists on the remote.

Onboarding, Windows setup, and the security model → Multi-machine docs.

Configuration

Set these under environment: in docker-compose.yml (all optional):

Variable	Default	Meaning
SAMPLE_INTERVAL	10	Seconds between samples
RETENTION_DAYS	180	How long history is kept
PRESSURE_FREE_MB	2048	Free VRAM below this counts as "pressure"
PORT	9800	Dashboard port
MCP_PORT	9810	Port for the built-in read-only MCP server
ENABLE_MCP	1	Set 0 to run the dashboard without the MCP server
WATCH_CONTAINERS	—	Extra containers to scan for OOM (comma-separated)
WATCH_SERVICES	—	systemd units to always show, even vendor ones (comma-separated)
CHECK_UPDATES	true	Set false to disable the daily GitHub-releases check (no outbound calls)

History lives in ./data/gpu.db (a bind mount), so it survives restarts and upgrades. Alerts, the systemd D-Bus mount, and per-server tuning → Configuration docs.

Under the hood

The hub stitches nvidia-smi, the Docker API, model-server APIs (Ollama, vLLM, llama.cpp, A1111, …), systemd D-Bus, and /proc + /sys into one sampled view, persisted to SQLite and downsampled on read so a six-month range loads as fast as the last hour. Single page, vendored Chart.js, no build step.

• 30+ recognised model servers → Model servers
• /metrics Prometheus endpoint + Grafana dashboard → Prometheus & Grafana
• The full data pipeline + caller attribution → How it works

Connect an AI agent (MCP)

Your homelab is now legible to AI agents — point a client at one URL and it can see every host, container, GPU and disk. Read-only, no extra setup.

HomeLab Monitor isn't just a dashboard for you anymore; it's context for your AI agent too. A read-only MCP server is built into the same container (served on :9810) — so Claude, Claude Code, or any MCP client connects in one line and explores your whole lab through 12 named tools, with the same coverage you see on the dashboard: hosts, containers, systemd services, GPU and who's driving it, per-process RAM, AI model servers, disk treemaps, history and alerts.

# the dashboard is on :9800; the MCP server rides along on :9810
claude mcp add --transport http homelab http://YOUR-HUB:9810/mcp

Once connected, skip the tab-hunting and just ask — the agent picks the right tools:

• "My GPU's been pinned for an hour — which model server is loaded, and who's actually calling it?"
• "What's eating /backup? Give me the biggest folders and flag anything that looks like runaway logs."
• "Which host is lowest on RAM right now, and what's the top process holding it?"
• "I want to reboot and run an OS upgrade this weekend — which box needs it most, and what's a safe order given what's running on each?"

Read-only by design — there are no write tools, so an agent can look but never touch your fleet. Turn it off anytime with ENABLE_MCP=0. Full tool list & setup → MCP docs.

Security

This is a host monitor: it runs with host access and a read-only Docker socket, root mount, and D-Bus socket — a broad footprint by design. Keep it behind your LAN/VPN/firewall and don't expose it to the public internet. Details → docs.

⭐ Support the project

If HomeLab Monitor saves you a browser tab or two, a ⭐ on GitHub genuinely helps other home-labbers find it. Thank you!

Contributing

Issues and PRs are very welcome — especially new model-server probes, new monitors, and GPU back-ends. This is a hobby tool meant to help fellow home-labbers, so be kind. See CONTRIBUTING.md.

License

MIT — see LICENSE.