−50%

50% off your first month on any paid plan · code WELCOME50 · until Nov 30

Start free →

Free tool · Cron

Cron Frequency Calculator

See how many times a schedule runs per hour, day, week, month and year.

Calculator free · no sign-up
Need to build an expression from scratch? Use the visual builder →
Per hour
—
Per day
—
Per week
—
Per month
—
Per year
—

Next 5 runs

A five-field expression hides a surprising number of runs

*/5 * * * * runs every five minutes. Obvious enough. But 0 */4 * * 1-5? That one fires 30 times a week, not 42, because weekends are excluded, and most people don't notice until the reporting job they expected to run on Saturday just... didn't.

Once you combine day-of-month with day-of-week, standard cron's OR logic kicks in and the schedule fires more often than the expression suggests. This calculator expands every field, counts the real total across a full year, and shows you the next scheduled runs so you can verify the pattern matches what you actually intended. If the expression itself looks wrong, the expression builder helps construct one visually, and the validator confirms the syntax before you deploy it.

How cron frequency is actually calculated

Each cron expression has five fields: minute, hour, day-of-month, month, and day-of-week. The calculator resolves every field into its set of matching values, then counts how many unique timestamps fall within each time window (per hour, per day, per week, per month, per year).

For minute and hour, the math is straightforward multiplication. */15 * * * * matches minutes 0, 15, 30, 45, so that's four values per hour and 96 runs per day. But things get tricky with the date fields. Months have different lengths (28, 29, 30, or 31 days), which means 0 0 * * * gives you 365 runs in a regular year and 366 in a leap year. And if you schedule something for the 31st, it only fires in seven months: January, March, May, July, August, October, December. The other five months quietly skip it.

The day-of-month and day-of-week OR rule

This trips up even experienced sysadmins. When both day-of-month and day-of-week are restricted (not *), standard cron treats them as an OR condition. 0 9 15 * 1 doesn't mean "the 15th if it's a Monday," it means "every 15th and every Monday." That's roughly 67 runs per year, not the 12 you'd expect. We've seen teams debug this for hours before realizing the schedule was working correctly, just not the way they read it.

Some cron implementations (Quartz, for example) use AND logic instead. If you're on Kubernetes CronJobs or a standard Linux crontab, it's OR. Always worth checking which flavour your system runs.

Cron expression syntax: the five fields

┌───────────── minute (0-59)
│ ┌─────────── hour (0-23)
│ │ ┌───────── day of month (1-31)
│ │ │ ┌─────── month (1-12)
│ │ │ │ ┌───── day of week (0-7, 0 and 7 = Sunday)
│ │ │ │ │
* * * * *

Each field accepts specific values, ranges, lists, and steps. Here's what the special characters do:

SymbolMeaningExample
*Any value* * * * * = every minute
,List of values0 8,17 * * * = 8 AM and 5 PM
-Range0 9 * * 1-5 = weekdays at 9 AM
/Step*/10 * * * * = every 10 minutes

Shortcuts like @daily (equals 0 0 * * *), @hourly (0 * * * *), and @weekly (0 0 * * 0) are supported by most cron implementations. The calculator above expands them automatically. For a full reference, the cron cheat sheet covers every symbol and shortcut.

What does each time window actually tell you?

Runs per hour matters for resource planning. A job that runs 12 times an hour competes with other scheduled tasks for CPU and I/O, and if one run takes longer than 5 minutes, the next invocation starts before the previous one finishes. That's where you get race conditions and duplicate rows in production databases.

The daily count is the number most ops teams think about first. A daily backup that actually runs 4 times a day wastes storage and network bandwidth. A health check that runs once when you expected it hourly leaves 23-hour blind spots.

Weekly and monthly totals reveal workday patterns. 0 8 * * 1-5 runs on weekdays, 5 times a week, 260-262 times a year (depending on which day January 1st falls on). If your SLA reporting assumes 365 daily data points but the cron only fires on weekdays, you'll have gaps in every weekend slot, and the dashboard will show zeroes that look like failures but aren't.

Then there's the yearly count, where edge cases pile up. Leap years add a day in February. A schedule set to 0 0 29 2 * runs once every four years. 0 0 1 */2 * (first day of every other month) fires 6 times, not 6.08 or any other averaged number. Months aren't evenly spaced, but cron doesn't care about that.

Common frequencies and their actual execution counts

A few expressions everyone writes, with the numbers that often surprise people:

ExpressionMeaningPer dayPer year
* * * * *Every minute1,440525,600
*/2 * * * *Every 2 minutes720262,800
*/5 * * * *Every 5 minutes288105,120
*/10 * * * *Every 10 minutes14452,560
*/15 * * * *Every 15 minutes9635,040
*/30 * * * *Every 30 minutes4817,520
0 * * * *Every hour248,760
0 */2 * * *Every 2 hours124,380
0 */4 * * *Every 4 hours62,190
0 */6 * * *Every 6 hours41,460
0 */12 * * *Every 12 hours2730
0 0 * * *Once a day (midnight)1365
0 8 * * *Once a day at 8 AM1365
0 8 * * 1-5Weekdays at 8 AM—~261
0 9,17 * * 1-59 AM and 5 PM on weekdays—~522
0 0 * * 0Once a week (Sunday)—52
0 0 * * 1Every Monday—52
0 0 1,15 * *1st and 15th of month—24
0 0 1 * *First of month—12
0 0 1 1 *Once a year (Jan 1)—1
0 0 29 2 *Feb 29 only (leap year)—~0.25

Numbers shift slightly in leap years and by how weekdays fall. The calculator above accounts for the actual calendar, not averages.

Check your frequency before it checks you

A job that fires 288 times a day has overlap problems. If the previous run hasn't finished when the next one starts, you get lock contention, duplicate inserts, or worse: a Stripe webhook processor that runs twice on the same event and charges a customer double. Frequent jobs need a flock guard or mutex, and you should know the count before deciding whether to add one. Honestly, the number of times we've seen a missing lock cause duplicate charges is a bit maddening.

Monthly jobs have the opposite problem. They fail silently and nobody catches it for 30 days. By the time someone notices the report is missing, a month of data is gone. Ran into this once with a quarterly compliance export: the S3 bucket policy changed, the upload started failing, and nobody checked for 90 days because the job "only runs four times a year." Infrequent jobs need someone (or something) watching, because they won't announce their own failure.

Edge cases the calculator handles

Cron expressions interact with the calendar in ways that aren't always obvious:

  • 0 0 29 2 * runs only in leap years. In non-leap years, February has 28 days and the job never triggers. Fine for a "leap year cleanup" script, but disastrous if you meant "end of February" and forgot that cron doesn't understand "last day of the month."
  • April, June, September, and November have 30 days, so a job scheduled for the 31st skips those four months entirely, plus February. You end up with 7 runs per year instead of 12. Have you checked whether your monthly job actually lands on a day that exists in every month?
  • The calculator shows UTC-based counts. If your cron daemon runs in a timezone with daylight saving transitions, one day a year has 23 hours and another has 25. A */30 * * * * job fires 46 times on the short day and 50 on the long one. Not huge, but it matters if you're counting events for billing or audit logs.
  • 0 0 1-30/7 * * fires on the 1st, 8th, 15th, 22nd, and 29th. Five runs, not four. Easy to misjudge when the range doesn't divide evenly by the step.

Theoretical schedules vs. what actually ran

This calculator tells you how many times a job should execute. Whether it actually did is a different question entirely. Servers reboot mid-run. Cron daemons freeze after an OOM kill. A job exits 0 even though the upload to S3 failed halfway through. The schedule says 365 runs per year, but production logs show 347, and those 18 missing runs are silent data gaps nobody noticed until quarter-end reporting.

Cron job monitoring bridges that gap. It tracks every execution, flags the ones that ran late or ran too long, and alerts when a scheduled run never shows up. The calculator gets you the expected number; monitoring tells you whether reality matched. You can pair both with the timezone converter if your team works across regions and needs to verify when each job fires in local time. If your SLA promises a certain uptime percentage, the uptime calculator shows exactly how many minutes of downtime that allows — context that matters when a missed cron run counts as an outage. For a deeper look at how cron syntax works, the complete cron jobs guide covers everything from basic expressions to logging strategies that make debugging missed runs less painful.

Frequently asked questions

It parses the five fields of a cron expression (minute, hour, day-of-month, month, day-of-week), expands each field into its set of matching values, then counts how many execution timestamps fall within each time window. The result is the number of times the job will run per hour, day, week, month, and year.

Most likely because both day-of-month and day-of-week are set to specific values. Standard cron (Linux crontab, Kubernetes CronJobs) treats these two fields as an OR condition, so the job fires when either condition is true. For example, "0 9 15 * 1" runs on every 15th AND every Monday, not just on Mondays that fall on the 15th.

No. Only seven months have 31 days (January, March, May, July, August, October, December). The job silently skips April, June, September, November, and February. If you need a true end-of-month schedule, most cron implementations don't support "last day" natively, so you'd use a wrapper script that checks the date.

A daily job (0 0 * * *) runs 366 times instead of 365 in a leap year. A job set to "0 0 29 2 *" runs once in a leap year and zero times in other years. The difference is small for most schedules, but matters for jobs tied to February 29 specifically.

Yes, if the cron daemon uses a timezone with DST transitions. On the spring-forward day, the 2:00-3:00 AM hour doesn't exist, so jobs scheduled in that window are skipped. On the fall-back day, the 1:00-2:00 AM hour happens twice, and whether the job runs once or twice depends on the cron implementation. UTC-based cron avoids this entirely.

More free tools

All 20 tools →

Turn a one-off check into a monitor.

feature
Pricing Docs
Log in Start free →