Cronicle

Last reviewed Sep 13, 2026 · 901 words

Cronicle is the best answer I know to "I have 40 cron jobs across 3 machines and no idea which ones ran last night." It gives every job a page with a run history, live streaming output, a success rate and a retry policy, and it does this on a single Node.js process using about 256 MB of RAM. What it is not, despite the Zapier comparison, is an integration platform. Cronicle runs your scripts on a schedule; it does not know what a Slack message or a Google Sheet is.

Crontab's real problem is visibility, and Cronicle fixes exactly that

A crontab entry fails silently. You find out three weeks later when the backup directory is empty. Cronicle captures stdout and stderr of every run, keeps the log, records the exit code and elapsed time, and emails or webhooks you on failure. Each job also gets a timeout, a maximum concurrent-run count and a "catch up" flag that decides whether missed runs during downtime are replayed. Those four settings cover most of what people bolt onto cron with wrapper scripts and flock.

The scheduling grammar is broader than cron too: you can pick specific years, months, weekdays, hours and minutes from a UI, or type a plain cron expression if you prefer. Time zones are per-event, which matters the day you move a server and every job shifts by an hour.

Multi-server is the feature that justifies it over Ofelia

One Cronicle install has a primary server and any number of worker servers. Workers join by sharing the same secret_key from conf/config.json and are grouped by hostname regex, so a job can be pinned to one box, to a group, or to "whichever member of the group is least loaded." The primary streams job output back to the UI in real time, and if the primary dies a worker with backup rights takes over.

If you only have one host and only Docker containers to schedule, Ofelia is smaller and lives entirely in compose labels. Cronicle earns its place once the second machine appears, or when a non-technical colleague needs a button that says "run now."

Install takes about 5 minutes on bare metal

The maintainer's install script pulls the release tarball into /opt/cronicle:

curl -s https://raw.githubusercontent.com/jhuckaby/Cronicle/master/bin/install.js | node
cd /opt/cronicle
bin/control.sh setup    # first server only; initialises storage
bin/control.sh start

The UI is on port 3012 and the first login is admin / admin; change it before you do anything else and put a reverse proxy with TLS in front, because the credentials and the job scripts travel in the clear otherwise. Storage defaults to plain JSON files under data/, which is fine for a homelab and trivially backed up with the rest of /opt/cronicle. S3 and Couchbase storage back ends exist for larger setups.

There is no first-party Docker image, only community ones. They work, but you are trusting a third party's Dockerfile with a tool that runs arbitrary shell as whatever user it is started as, so read the Dockerfile or build your own from the install script.

Jobs are shell by default, and that is most of the point

The bundled Shell Script plugin is what nearly everyone uses: paste a script into the event, pick an interpreter, done. An HTTP Request plugin covers "hit this URL every 10 minutes" without a script. Beyond that you write plugins as any executable that reads JSON on stdin and writes JSON on stdout, which is simple enough that people wrap Python and Go tools in an afternoon.

Chaining is the feature that quietly replaces a lot of Makefile logic: an event can trigger another event on success or on failure, passing its data along. A nightly database dump, then an upload, then a prune, each with its own log and its own retry policy, is 3 events and 2 chain links.

Where it is the wrong tool

Cronicle is a scheduler for administrators. If the thing you want to automate is "when a form is submitted, add a row and send an email," you want n8n or another node in the automation category, not this. If you need DAGs with hundreds of interdependent tasks, backfills and data lineage, that is Airflow territory and Cronicle's chaining will feel thin. And if your only concern is "did my cron job run," Healthchecks does that as a dead-man's switch with no scheduler at all, and pairs well with Cronicle rather than competing with it.

The community is small: 5,800 stars, one primary maintainer, and releases arrive when they arrive. The code is stable and the MIT licence means you can patch it yourself, but do not expect a plugin marketplace.

What I'd do

Install it bare-metal on the box that already runs your scripts, front it with Caddy, change the admin password, and move your crontab over one entry at a time so you can compare logs for a week. Add a worker the moment a second host needs scheduled work. Keep Healthchecks pinging from your most important jobs as a second opinion. For anything that involves APIs, webhooks and business logic rather than scripts, leave Cronicle alone and reach for n8n.