Skip to content

Batch rules

The batch rules cover Job and CronJob resources, diagnosing failures and misconfigurations that prevent scheduled or one-off workloads from completing successfully.

Job rules

job/failed

Severity: error | Confidence: 0.85 | Applies to: Job

Detects Jobs whose Failed condition is True. The rule tailors its explanation to the specific failure reason — backoff limit exhaustion, active deadline exceeded, or a general failure — and attaches container log evidence from owned pods where available.

When it fires

The Job's status.conditions contain a Failed condition with status True.

Failure reasons and their explanations:

Reason Title Explanation
BackoffLimitExceeded Job exceeded its backoff limit The Job retried its pods up to backoffLimit and gave up. The container is failing on every attempt.
DeadlineExceeded Job exceeded its active deadline The Job ran longer than activeDeadlineSeconds and was terminated before completing.
(other) Job failed The Job did not complete successfully. Inspect the failed pods.

Example finding

Job exceeded its backoff limit The Job retried its pods up to backoffLimit and gave up. The container is failing on every attempt.

Remediation

# Inspect failed Job pods and their logs
kubectl logs job/<name> -n <namespace>
kubectl describe job <name> -n <namespace>

CronJob rules

cronjob/suspended

Severity: info | Confidence: 0.90 | Applies to: CronJob

Flags CronJobs that have spec.suspend: true, which prevents any new Jobs from being created. This state may be intentional (for example during maintenance) but is surfaced as an info finding so it is visible in diagnoses.

When it fires

spec.suspend is true.

Example finding

CronJob is suspended The CronJob is suspended, so no new Jobs are created until it is resumed. This may be intentional.

Remediation

If the CronJob should be running:

kubectl patch cronjob <name> -n <namespace> -p '{"spec":{"suspend":false}}'

cronjob/job-failures

Severity: error | Confidence: 0.75 | Applies to: CronJob

Detects CronJobs whose most recently owned Jobs are in a failed state. The scheduled task is not completing successfully.

When it fires

One or more Jobs owned by the CronJob (visible in the resource graph) have a Failed condition set to True.

Example finding

CronJob has 2 failing Job(s) Jobs created by this CronJob are failing, so the scheduled task is not completing successfully.

Remediation

# List Jobs created by the CronJob, sorted by creation time
kubectl get jobs -n <namespace> --sort-by=.metadata.creationTimestamp

# Inspect the most recent failed Job
kubectl describe job <job-name> -n <namespace>
kubectl logs job/<job-name> -n <namespace>