We had performance issues in South Central US due to SQL database upgrade on two databases. Users might have noticed intermittent connection issues or 500's during the impact time. We worked with SQL Azure team and fixed the database upgrade issues.
We lost the database quorum during the upgrade and that caused all the write operations to fail on the impacted databases.
We are actively investigating root cause on why the database quorum was lost during upgrade.
Incident Timeline: 2018/04/13 21:35 UTC to 2018/04/14 00:55 UTC