Session: 2026-06-04
Cluster Janitor — Jobs running >2h scan
Scan time: ~15:00 UTC
Scope: All namespaces, jobs in Active state with duration >2 hours
Findings
| Namespace | Job Name | Duration | Pod Status | Root Cause | Severity | Action |
|---|---|---|---|---|---|---|
| default | legend-smart-digest-29675640 | 5h19m | ContainerCreating (stuck) | Missing ConfigMap legend-smart-digest-script in default namespace. The ConfigMap exists in the legend namespace but not where the CronJob references it. Pod repeatedly fails to mount volume script. Backoff limit (6) exhausted. | COSMETIC — non-critical cron job, no downstream impact; only blocks a dead pod and wastes cluster resources | Deleted job default/legend-smart-digest-29675640 via kubectl. CronJob has concurrencyPolicy: Forbid, so deletion is safe — next schedule cycle will create a fresh Job (which will also fail unless the ConfigMap issue is fixed). |
Summary
- 1 stale job found and remediated
- No blocking/degraded jobs discovered
- Root cause is mis-namespaced ConfigMap reference in CronJob spec (
defaultvslegend) - Escalation not filed: cosmetic severity, low risk. If the issue persists across schedule cycles, open a child task to fix the CronJob spec to reference the correct namespace for the ConfigMap.