Skip to content
Uptime

Start here

Overview

On this page

Uptime is self-hosted uptime and API monitoring. It checks your sites and APIs on a schedule, tells you when a failure is confirmed, and never calls a missing check healthy. You run it yourself, it is free and open source under the MIT license, and the source is at https://github.com/just-every/run-uptime.

How it works

Uptime has four parts, and it helps to know which does what.

  • A Cloudflare Worker serves the console and its APIs, and decides what each result means.
  • A D1 database stores every scheduled check, every attempt and every incident, so history is never rewritten.
  • A Cron trigger runs every minute to expire abandoned work and clear out old sign-in records.
  • A private runner, a small Node process that you start and supervise separately, makes the actual requests to your targets and sends back signed results.

You work in the console at /app once you have signed in. Only the installation owner can open it.

At the app origin, / is the Lobby, /setup is first-run setup, and /docs is public documentation. The separate marketing site is https://run-uptime.com/, with its /run page.

Because the runner is a separate process, a healthy Worker does not tell you the runner is working. The runner page shows how to keep it running.

What Unknown means

Every green mark in Uptime is a check that ran and passed. If a result never arrives, or arrives too late to count, Uptime shows that time as Unknown. Unknown is not healthy, and it is not failing either: it means Uptime has no fresh evidence either way.

Read more in Monitors and evidence.

Four kinds of monitor

  • HTTP(S) checks a URL and tests the response against your own success rules: status, body text, headers, JSON values and latency.
  • Keyword is an HTTP(S) check that requires a piece of text in the response body.
  • Incoming heartbeat waits for your job or service to call Uptime on schedule, which suits cron jobs and background workers.
  • Browser flow checks one bounded path in a real browser on a runner with a browser. It reports only the steps and outcomes you configured; it does not represent the rest of the site. See What Uptime does not do.

Notifications

When a failure is confirmed, Uptime can notify you by email, a signed JSON webhook or a Slack incoming webhook. Destinations apply to every monitor in the installation. See Notifications.

Who it suits

Uptime suits you if you want to run your own monitoring and read the evidence behind every alert. It fits a small team or one person who is comfortable running a Cloudflare Worker and a small Node process, and who would rather see an honest "no check received" than a reassuring green.

It does not cover everything. Uptime uses sampled evidence from the assigned runner networks, not a promise about every visitor's experience. See What Uptime does not do.

Where to go next

  1. Install Uptime on your own computer to look around.
  2. Deploy to Cloudflare when you want it running for real.
  3. Start the private runner so checks begin.
  4. Add monitors and notifications.