7 Open-Source Website Monitoring Tools

Uptime Kuma is this category's default answer, with 93.2k GitHub stars: one Docker command brings up a dashboard that checks HTTP(s), keyword and JSON query checks, raw TCP, ping, DNS records, WebSocket, push-based monitors, Steam game servers and Docker containers, with alerts through 90-plus notification services and multiple status pages, mappable to specific domains.

It is MIT licensed and runs on Docker or Node.js. The structural limit is that everything happens on the one machine you installed it on: the checks, the history and the alerting all share that box's fate. Our Uptime Kuma guide covers setup, the limits and the comparisons in depth.

Good fit: anyone with a Docker host who wants monitoring running before dinner.

Gatus takes the opposite interface philosophy from Uptime Kuma: the whole setup lives in YAML files, so your monitoring is diffable and sits in git next to the services it watches. Checks are written as one-line conditions on the status code, the body and the response time, covering HTTP, ping, TCP, DNS, TLS and domain expiry.

It ships as a Docker image, Helm chart or single Go binary, under the Apache 2 license, with roughly 40 alerting providers.

Good fit: teams that keep infrastructure in code review and want monitoring held to the same standard.

Upptime removes the server from self-hosting: checks run as GitHub Actions workflows every 5 minutes, an outage opens a GitHub issue that closes itself on recovery, and the status page is a static site on GitHub Pages. Response-time history lives in the repository itself.

The ceiling is built in. Checks come from GitHub's runners on a schedule you do not fully control, the interval floor is those 5 minutes, and alerting reaches Slack, Telegram, custom webhooks and more.

Good fit: open-source projects and portfolios that live on GitHub anyway and want a public status page for nothing.

Website checks in Prometheus go through the blackbox exporter, which probes endpoints from the outside over HTTP, HTTPS, DNS, TCP, ICMP and gRPC, exposing probe_success and timing metrics. One exporter probes many endpoints, and alerting rules fire on probe_success going to zero.

This is the heaviest option on the page for watching a website alone, and the strongest when the same stack already collects your application and server metrics. The monitoring arrives as parts, with Grafana and an alert router assembled around it.

Good fit: teams already running Prometheus who want uptime probes in the same query language as everything else.

Zabbix is a full monitoring platform under the AGPL, with website checks as one feature among servers, networks and applications. Nothing sits behind a paid tier; the company sells technical support instead of the software.

It is the long-haul choice here: a server, a database and a web interface of its own, and a learning curve to match. Running it for website checks alone means operating far more platform than the job needs.

Good fit: an organisation that wants one self-hosted platform for everything and can staff it.

Checkmk's free Community edition is open source and arrives more assembled than most of this list: discovery, agents and a working interface from the first boot, easily scalable up to about 100 hosts.

Its centre of gravity is infrastructure rather than websites, and the paid editions above Community are where synthetic monitoring lives. As the free self-hosted base for a server fleet that also wants HTTP checks, it earns its place.

Good fit: a fleet that needs server monitoring first and website checks second, from one free install.

OneUptime is the open-source attempt at the whole category in one platform: monitors, status pages, incident management, on-call and logs. Self-hosting it gives you all of that without a bill.

Breadth is also the cost. You are operating a platform, with the moving parts that implies, where Uptime Kuma or Gatus would have been one container. The hosted version exists for teams that want the breadth without the operations; its free tier is mostly a status page with monitors billed per active monitor.

Good fit: a team that wants status pages, on-call and monitoring from one self-hosted codebase and accepts the operational weight.