A routine database migration caused an unexpected database lock, triggering an outage from 09:50 PT to about 10:00 PT. The process was terminated restoring service. We identified a misbehaving process that led to the lock and have addressed this issue.