We seem to have intermittent cronjobs failures. A little less than one per day out of probably 100, some of which are run every 5 minutes. So not often really.
The oddness is that when I go to check it out, there is no pod for this specific failure. It is set to keep the last 2 failures, and for code problems or what not it does. But for this specific failure the pod is never there. The job is present, and says the pod failed as well.
The event in k8s events is
Back-off pulling image "foo"
But the next run of the cronjob reports the same image is already present. So clearly it actually did pull the image down. Which is why I would really love to see the pod logs and describe to see if there is more info on the error.
So anyone have any idea why the pod is gone?