Errors / checked October 9, 2026
too many missed start times. Set or decrease .spec.startingDeadlineSeconds or check clock skew
The error
A TooManyMissedTimes Warning event on a Kubernetes CronJob (kubectl describe cronjob or kubectl get events), usually after the controller was down, the CronJob was suspended, or the cluster clock jumped.
Quick fix
- Set .spec.startingDeadlineSeconds (for example 200). The controller then only counts missed runs inside that window, so a long outage no longer counts as hundreds of misses.
- Check clock skew on the control plane nodes; a clock that jumps can make many schedules look missed at once.
- If the CronJob was suspended on purpose, expect this once on resume; it still starts one Job for the most recent missed time.
Why this happens
Every 10 seconds the CronJob controller counts how many scheduled times it missed since the last run. Over 100 and it records this warning, starts one Job for the latest missed time and never starts the earlier ones.
Source: Kubernetes: CronJob (job creation, missed schedules)
Other solutions
- Make the job safe to repeat: Kubernetes documents that a CronJob can create two Jobs for one time, or none.
- If the job only calls a URL, a scheduler outside the cluster keeps running during control plane outages and logs each run. RunLater calls the URL on your cron schedule, retries failures and keeps each response 7 to 90 days.
Check a schedule before you deploy it: the cron expression explainer shows the next 10 run times.
Frequently asked questions
Will the missed runs be started later?
No. The controller starts one Job for the most recent missed time; the earlier ones are skipped however many there are.
What value should startingDeadlineSeconds have?
Shorter than the gap between runs and long enough to cover a brief outage, for example 200 seconds for a job that runs every few minutes. Under 10 seconds the CronJob may not be scheduled at all, since the controller checks every 10 seconds.
Does suspending a CronJob cause this?
It can. Runs scheduled while a CronJob is suspended count as missed, so a long suspension can push the count past 100.
Have the URL called on schedule, with retries
RunLater calls your URL on a cron schedule or at an exact time, in your timezone, retries failed calls with backoff and logs each response, so a timeout or a 5xx shows up in a log instead of disappearing.