The service has gone down from 7:00 to 7:22 am (Greenwich Mean Time) due to a deadlock on one of the db clusters. The db service has been restarted and service is now back to normal. We are currently investigating on the detailed issue to prevent it and to develop possible auto-recovery measures.