One of our databases shards failed over to a standby, but application processes did not come back online cleanly. We have identified and restarted the affected processes.