Cronbell is an open-source tool for monitoring cron jobs, scheduled tasks, and background workers, written in any programming language.

Problem: if a job does not run, it does not throw an error, nor does it log anything. The backup job stopped three weeks ago, and the invoice job failed after one deploy, usually only discovered when customers report it.

For example: the database backup job runs at 2 AM, allowing a 30-minute delay. Any night the job does not run, Cronbell will send a message via Telegram or Slack to you around 2:30 AM, along with the last successful run. If the job runs but encounters an error, the message will also include the exit code and the last output snippet, so you can understand why.

How it works: each time a job runs, it sends an HTTP request (just add a line curl at the end of the crontab). Cronbell knows the schedule and time zone of the job, so when a run is overdue, exits with an error code, or runs past the allowed time, it sends alerts via email, Telegram, Slack, Discord, ntfy, or webhook (you can choose a specific channel for each job), and reports back when the job is running normally again. The dashboard displays a 30-day status range for all jobs, uptime for 7 and 30 days, a history of each failure with which channel reported it, each run with time and the last 10 KB of output, as well as sample code for crontab, shell, Kubernetes, and Spring Boot.

For: backend developers, DevOps, anyone running their own server with scheduled jobs (backups, data synchronization, sending reports, cleaning up).

Getting started: self-hosting with a command docker run, data stored in SQLite (PostgreSQL option), with instructions for setting up on your own server with HTTPS. If you prefer not to self-host, you can request a link to use the pre-running beta at cron.steelreed.com by contacting [email protected]. License AGPL-3.0, free of charge.

Difference: Cronitor or Dead Man's Snitch are paid services with data hosted on their servers; Cronbell is self-hosted and free on your machine. Healthchecks.io is also open-source and very good; Cronbell uses the same type of ping URL (UUID) so current scripts just need to change the host and key, and add specific features: runs compactly in a container with SQLite, the dashboard has a 30-day status range and a history of each failure, and a starter for Spring Boot (available on Maven Central) automatically finds jobs @Scheduled, Quartz, ShedLock to report each run without requiring code modifications. Cronbell is developed by a person in Vietnam (with AI assistance), under the name Steelreed.

Source code: https://github.com/steelreed/cronbell