For plain host and container health, Beszel gives you the dashboard and alerts without Datadog's per-host bill or a 500MB agent.
Datadog is the industry standard for observability, and it prices like it — $15 to $23 per host per month before custom metrics and log ingestion push the bill higher. A surprising share of teams pay that to watch a handful of servers for CPU, memory, disk and uptime. If that describes you, Beszel delivers exactly that view from an agent that uses under 10MB of RAM, against Datadog’s 200–500MB, with no per-host meter at all.
Out of the box Beszel tracks the metrics most small fleets actually alert on: CPU, memory, disk and network usage, plus per-container stats for anything running under Docker. It ships as a single hub binary with a clean web dashboard and lightweight agents that report in, and alerting over email, Discord, Telegram and webhooks is built in rather than bolted on. For a homelab, a few VPS boxes, or a startup’s production handful, it covers the daily question — is everything healthy — without ceremony.
The trade-off is scope. Beszel is infrastructure monitoring, not full observability: there is no application performance monitoring, no distributed tracing, and no log aggregation. Datadog’s strength is correlating a slow request across services and drilling into a flame graph, and Beszel does none of that. Its dashboards are deliberately simple. If you need to debug why one endpoint is slow under load, Datadog earns its price; if you need to know that a box is running out of disk, Beszel does it for free.
Standing it up is an afternoon. Deploy the Beszel hub with Docker, run the install one-liner on each server to register an agent, and watch the systems populate the dashboard within minutes. Wire your alert channel, set thresholds, and you can decommission the Datadog agents host by host as you gain confidence. Because the agents are so small, there is little risk in running both side by side for a week before you cancel the subscription.