Beszel
Lightweight server monitoring with Docker stats, history and alerts
Beszel is a small hub and agent pair that tracks CPU, memory, disk, network and per container usage across your machines, with historical graphs and alerts. It is MIT licensed and much simpler to set up than a full Prometheus and Grafana stack.
Best for when
- you want host and container graphs without building a monitoring stack
- you watch a handful of machines, not hundreds
- you want simple alerts for offline hosts and full disks
Watch out for
- it tracks a fixed set of system metrics, not your applications' own metrics
- the agent needs Docker socket access for container stats, which is a privileged permission
- it is a younger project than Prometheus, with a smaller community and fewer integrations
Enough monitoring for a home lab
You run the hub once, which provides the web interface, and install a small agent on each machine you want to watch. The hub shows each system with its load, memory, disk and temperature, plus per container CPU and memory, and keeps history so you can see what happened overnight.
Alerts cover the common cases, such as a host going offline or a disk filling up, through email or many notification services. It suits people with a few machines who want useful graphs in ten minutes and do not want to learn PromQL.
What running it involves
Both parts are small Go binaries or containers with a low memory footprint, and the hub stores its data in an embedded database built on PocketBase. Back up the hub's data folder and you have kept everything.
The agent reads system stats and, for container metrics, the Docker socket, so treat it as privileged. Keep the hub behind a VPN or a trusted login. The flip side of simplicity is depth: there are no custom queries or application metrics, so it complements rather than replaces a full observability stack.
