We believe the issue is related to high CPU usage in our backend database and are failing over to a replica. We'll update again at 6:15 PST.