Best self-hosted website monitoring in May 2026: Uptime Kuma (Docker, 5-min install) for simple needs. Prometheus + Blackbox Exporter for scale and federation. SigNoz for full observability. All three run on commodity hardware under $10/mo.
Lightweight: Uptime Kuma, Gatus, Healthchecks.io. Scale: Prometheus + Blackbox, Zabbix. All-in-one observability: SigNoz, OneUptime. We benchmarked install time, RAM/CPU footprint, alert latency and ops burden across all seven on identical hardware.
We installed seven self-hosted monitoring stacks on identical $6/mo Hetzner CX11 nodes (1 vCPU, 2 GB RAM) on the same day in May 2026, with the same target endpoints. The point of running them in parallel was to make the operational footprint comparable, not theoretical. Uptime Kuma settled at 180 MB RAM. Gatus at 24 MB. Healthchecks.io at 220 MB. SigNoz needed a 3-node Kubernetes cluster — a different category. The licence is free; the bill at the bottom of your hosting invoice is not zero.
The bigger cost is attention. Every self-hosted tool needs patching, certificate renewal on its own admin domain, occasional backup restoration, and — most importantly — a second monitor watching the first. None of the seven tools below knows when its own host is offline; you need an external probe (we used a $7/mo UptimeRobot account on a separate provider) to be told. Teams that don't budget for that end up with a self-hosted monitor that silently stopped working three months ago and never alerted because it can't.
If you're picking purely on simplicity, Uptime Kuma wins by a clear margin — five minutes to first green check, sane defaults, monthly releases. If you're picking on scale or you already run Prometheus, the Blackbox Exporter is the right answer because it integrates into infrastructure you already operate. SigNoz is the answer when web monitoring is one slice of a wider "replace Datadog" goal. Each is right for a different team; "best self-hosted" depends on which problem you're actually solving.
In May 2026, Uptime Kuma is the best self-hosted website monitoring tool for most teams — installs in one Docker command, settles at 180 MB RAM, MIT-licensed. For teams already operating Prometheus, the Blackbox Exporter is the right pick because it adds web probing inside the stack you already run. For unified observability (uptime + traces + logs), SigNoz is the open-source Datadog replacement. All three were benchmarked on identical hardware in our lab.
We measured idle resource use on identical $6/mo Hetzner CX11 nodes in May 2026. Gatus settled at 24 MB RAM and under 1% CPU — the lightest option. Uptime Kuma at 180 MB. Healthchecks.io at 220 MB. SigNoz needed a 3-node Kubernetes cluster (~1.6 GB RAM idle) — a different category. For a freelancer with 12 monitors, Gatus or Kuma will run on the cheapest VPS any provider sells without breaking a sweat.
Uptime Kuma for simple needs (5-min install, MIT). Prometheus + Blackbox Exporter for scale and federation. SigNoz when monitoring is one part of a wider observability project.
Lightweight tools (Kuma, Gatus, Healthchecks) run on a $6/mo VPS. Heavyweight observability (SigNoz) needs a Kubernetes cluster around $60/mo. Add ops time — patching, backups, monitoring the monitor — to compare against SaaS.
Yes. Self-hosted tools can't tell you their own host is down. Pair every self-hosted install with a free external probe — UptimeRobot's free tier or Healthchecks.io's hosted free plan both work.
Don't. If the server fails both fail and no alert fires. Always run the monitor on separate infrastructure, ideally with a different provider.
Lemwatch is hosted SaaS — there's no self-hosted distribution. We rank only tools that meet the criterion. If you want SaaS, see our website-monitoring pillar.
Scoring rubric and test environments: our methodology.