Errors / checked October 9, 2026
Job has reached the specified backoff limit
The error
A Kubernetes Job (often one a CronJob created) whose status condition is type Failed with reason BackoffLimitExceeded. kubectl describe job shows the same text as a Warning event from job-controller.
Quick fix
- Find out why the pods failed first: kubectl get pods --selector=job-name=<job> lists them, and kubectl logs <pod> (add --previous for a restarted container) shows the output. kubectl describe pod shows the exit code and reasons such as OOMKilled.
- Fix that cause (a crashing command, a missing secret or config value, a memory limit, an unreachable service), then create the Job again. A failed Job does not restart on its own; a CronJob starts a new Job at its next scheduled time.
- Raise .spec.backoffLimit only if the failures are transient, such as a dependency that is briefly down. The default is 6.
Why this happens
The Job controller counts pod failures. When the count reaches .spec.backoffLimit, it marks the Job Failed with this message and terminates any pods still running. With restartPolicy Never, each failed pod counts once. With restartPolicy OnFailure, the container restarts in pending and running pods count instead, and the pod is terminated when the limit is reached, which can take its logs with it.
Other solutions
- Set restartPolicy: Never while debugging, so each failed attempt leaves a pod whose logs you can read.
- Use .spec.podFailurePolicy to fail the Job at once on an exit code that retrying cannot fix, or to ignore pod disruptions so they do not count toward the limit.
- A similar failure with reason DeadlineExceeded reads "Job was active longer than specified deadline". That is .spec.activeDeadlineSeconds, which takes precedence over backoffLimit.
- If the job only calls a URL, a scheduler outside the cluster can make the request on your cron schedule instead. RunLater retries a failed call up to 10 times with exponential backoff, keeps each response 7 to 90 days by plan, and emails you when a run fails after its retries.
Check a schedule before you deploy it: the cron expression explainer shows the next 10 run times.
Frequently asked questions
What is the default backoffLimit?
6. If .spec.backoffLimitPerIndex is set on an Indexed Job, backoffLimit defaults to 2147483647 instead and the per-index limit applies.
How long does Kubernetes wait between retries?
Failed pods are recreated with an exponential back-off delay of 10 seconds, 20 seconds, 40 seconds and so on, capped at ten minutes.
Where did the pod logs go?
With restartPolicy OnFailure the pod is terminated when the limit is reached, so its logs can be lost. With restartPolicy Never the failed pods stay until the Job is deleted or cleaned up by ttlSecondsAfterFinished or the CronJob history limits.
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.