We missed this out last time. It's possible in an HA scenario for a second pod to start later while the other job is in-progress. In that case, it would schedule two jobs. https://mattermost.atlassian.net/browse/MM-53747 ```release-note NONE ```