Backups that silently stopped. Reports that never sent. Heartbeat watches every scheduled job and alerts you by email, Slack or Discord the minute one misses its check-in.
No agents to install, no SDK, no config files. If it can make an HTTP request, it can be monitored.
Tell us how often the job runs, like every hour or every night, and how much slack to allow.
Every 1 day · 1 hour grace
Add a request to your unique URL at the end of the job. Works from bash, PowerShell, Python, Node and more.
backup.sh && curl …/ping/9f2c
If a ping is late, we alert you once, then again when the job recovers. No alert storms.
[DOWN] "backup" missed its check-inBuilt for developers, sysadmins and small teams who'd rather find out from us than from their customers.
Checked every minute. Late by even one grace period? You'll know.
Jobs can report their own errors with /fail for an instant alert, with the error output attached.
Ping /start first to track how long runs take, with median and p95. Get alerted when a run is slow or hangs.
Send alerts where your team already is. Webhooks plug into anything else.
Show live job status in your README or internal wiki with an embeddable SVG badge.
Every ping is logged. POST your job's output and read it right next to the failure.
Use a real cron expression in any time zone, DST included, or a simple interval. Alert after N failures in a row to skip one-off flakes.
No server to run cron on? Heartbeat calls your URL on schedule, with your own method, headers and body, and records every response.
Define checks in .heartbeat.yml and sync them from GitHub Actions. Trigger workflows on a schedule, or run jobs from VS Code with results reported automatically.
Heartbeat is in an early Pilot. There's no paid plan and no credit card. Your feedback decides what gets built next.
For side projects, personal servers and small teams
Optional, and always appreciated
Heartbeat is built and run independently. If it's useful to you, a donation helps cover hosting and fund the features users ask for. Telling us what you'd like to see helps just as much. See where donations go →
♥ DonateDonations open soon.
Heartbeat is an independent project. Here is exactly what your support pays for, in order.
Servers, database, email delivery and the domain. These come first, so monitoring stays reliable and free for everyone in the Pilot.
Development time goes to what Pilot users request most: new integrations, longer history, team accounts and more. Your feedback sets the order.
Our goal is $100,000 in donations. Reaching it lets us set Heartbeat up as a proper business with dedicated support, so it's here for the long run. Paid plans will be shaped by Pilot feedback.
Donations are voluntary gifts. They aren't tax-deductible and don't buy equity or extra features, and Pilot accounts get the same features whether or not you donate. Payments are processed by Ko-fi.
Donations open soon.
Anything that runs on a schedule and can make an HTTPS request: cron jobs, Windows Task Scheduler tasks, Kubernetes CronJobs, GitHub Actions schedules, backup scripts, ETL pipelines, serverless functions, and more. Every connection is encrypted: pings and the API are HTTPS only (plain HTTP is refused), and the URLs Heartbeat calls for you must be HTTPS too.
We check every minute. You're alerted within about a minute of a check passing its expected time plus the grace period you chose, or immediately if your job pings /fail.
No. You get one alert when a check goes down and one when it recovers. Nothing in between.
Yes. Treat it like a password: anyone with it can send pings for that check. Status badges use a separate read-only link, so they're safe to make public.
No. There's no agent or SDK. One curl, Invoke-RestMethod or HTTP call at the end of your job is all it takes.
Set up your first check in under a minute. Free during the Pilot.