Cron every minute: the expression, explained, in Linux crontab
The cron expression for “every minute” is * * * * *. This page shows what each field does, the next five times it will fire, the mistake most people make with it, and a working Linux crontab example you can paste.
The expression
| Field | Value | Meaning |
|---|---|---|
| minute | * | every minute |
| hour | * | every hour |
| day of month | * | every day of month |
| month | * | every month |
| day of week | * | every day of week |
Next five runs (UTC, from now)
- 2026-09-17 22:02 UTC
- 2026-09-17 22:03 UTC
- 2026-09-17 22:04 UTC
- 2026-09-17 22:05 UTC
- 2026-09-17 22:06 UTC
Mistakes people make with this schedule
- Running a slow job every minute: if one run takes 90 seconds, two overlap. Use a lock, or a longer interval.
- Assuming the server clock is your clock. Cron runs in the machine’s timezone unless the runtime says otherwise; the examples here are UTC.
- No lock around a job that might overlap itself, and no alert when it silently stops. That second one is what Supercrontab is for.
Run it in your stack
# crontab -e * * * * * /usr/bin/curl -fsS https://example.com/cron/run > /dev/null
Lines run under the user’s crontab in the system timezone. Redirect output or cron will email it to you.
Questions
Does * * * * * run in my local time?
Only if the scheduler is configured for it. crontab uses the system timezone; Vercel and GitHub Actions always use UTC; Laravel and node-cron take a timezone option. In Supercrontab the timezone is a field on the job.
What if the previous run is still going?
Classic cron starts the next run anyway. Add a lock (flock on Linux, withoutOverlapping() in Laravel) or let Supercrontab skip the tick while the previous request is open.
Is there a shorter way to write this?
No. The five-field form is the shortest portable version.