Views:

Manually invoke the collector, for example to recover from a missed scheduled run.

Procedure

  1. Create a one-off Job from the CronJob's current template:
    CRON=$(kubectl -n "$NS" get cronjob -o jsonpath='{.items[0].metadata.name}')
kubectl -n "$NS" create job --from=cronjob/"$CRON" manual-run

kubectl -n "$NS" wait --for=condition=complete job/manual-run --timeout=300s
kubectl -n "$NS" logs --selector=job-name=manual-run --tail=200
    Important
    Important
    A manual run uses the CronJob's current schedule and scan-window settings — it does not accept a custom time window. If the resulting scan overlaps with a previous or upcoming scheduled run, chats in the overlapping period may be scanned twice.
  2. Clean up the manual job when you're done reviewing its logs:
    kubectl -n "$NS" delete job manual-run