HomeLab Monitor

HomeLab Monitor puts host health, Docker containers, systemd services, disks, network activity, uptime checks, and GPU usage on one self hosted dashboard. It is especially well suited to local AI servers because it can identify which process, container, or model is consuming GPU memory. NVIDIA, AMD, and Intel hardware are supported, although the available measurements depend on the operating system and driver.

You can install it with Docker Compose and keep historical metrics in a local SQLite database. Other Linux, Raspberry Pi, and Windows machines can be monitored over SSH without installing a permanent agent. The dashboard also detects model servers such as Ollama, vLLM, llama.cpp, ComfyUI, and TGI. Its read only MCP server lets compatible AI assistants inspect host, container, disk, and GPU information.

The convenience comes with an important security consideration. The default container receives extensive host access, including a writable Docker socket, systemd access, host networking, and process visibility. Container controls and automatic updates are enabled by default, while the dashboard has no built in authentication. Keep it on a trusted LAN or VPN. A supplied read only Compose override disables control actions and is the better starting point for most installations.

HomeLab Monitor suits smaller labs and local AI rigs that need practical monitoring without operating a separate Prometheus and Grafana stack. It is free under the MIT license, actively maintained, and still young enough that cautious deployment is sensible.