From about 3:21 pm eastern time to 3:34pm, the Dashboard and API were temporarily down due to a long-running database migration that was holding an exclusive lock. The long-running transaction was manually killed early to free up waiting transactions. Service resumed after said transaction was killed.