Introduction
Monitoring for jobs that are supposed to run
Alwaysbeat is a dead-man's-switch for scheduled work. Your job sends an HTTP request to a unique URL each time it finishes. If that request is late — or never comes — Alwaysbeat notifies you.
Cron jobs, backups, data pipelines, and background workers usually fail silently. The server is down, the crontab entry was deleted, a network mount vanished — and the first you hear of it is when someone asks where last night's backup went. Alwaysbeat closes that gap by watching for the absence of a signal rather than the presence of an error.
How it works
- Create a check with a schedule (how often the job should run) and a grace period (how long you'll tolerate it being late).
- Ping the check's URL from your job when it finishes successfully — for
example by appending
curlto your cron command. - Alwaysbeat computes the next deadline from the schedule. A background sweeper runs every minute; if a check is past its deadline plus grace, it flips to down and you get an alert.
- The next successful ping flips the check back to up and sends a recovery notification.
The key idea
You are not monitoring whether your job reports an error. You are monitoring whether it
checks in on time. A job that crashes before it can report anything is exactly the
case a dead-man's-switch catches.
Core concepts
| Term | Meaning |
|---|---|
| Check | One monitored job. Has a unique ping URL, a schedule, a grace period, and a list of notification channels. |
| Ping | An HTTP request your job makes to its URL. A plain ping means "I finished successfully." |
| Schedule | How often the job is expected to run — a simple interval (every 1h) or a cron expression in a specific timezone. |
| Grace | Extra time allowed after the deadline before the check is considered down. Absorbs normal variation in run time and start time. |
| Status | new (no ping yet), up (pinged on time), down (a ping is overdue), or paused. |
| Channel | Where alerts go: an email address, a Telegram chat, or a signed webhook. |
Where things live
| URL | Purpose |
|---|---|
app.alwaysbeat.com | Dashboard — create and manage checks, view history, manage API keys. |
ping.alwaysbeat.com | Ping endpoint — where your jobs check in. |
api.alwaysbeat.com | HTTP API — programmatic access, used by scripts and the Terraform provider. |
Next steps
Ready to set up your first check? The Quick start walks through it end to end in a couple of minutes. To understand the ping URL in depth, see Signaling jobs.