Several databases upgrade took longer than planned, which resulted in longer errors window for some customers. Now service is fully functional.
One database upgrade is still running, failover will take <1 minute and is expected to happen during nearest 40 minutes.
Last database failover was completed 9:32 UTC.