Service Maintenance, version 1710 (9.1.0.6820) release caused over 50GB consumption of DB storage.

In the Oceania region, 22 July 2019 (NZ) Microsoft released a Service Maintenance, version 1710 (9.1.0.6820) with a feature that caused over 50GB consumption of DB storage.

We noticed the spike in DB and raised a Microsoft ticket 119072222002126.

Microsoft Engineers advised that indexes were auto-generated by a new feature causing the increase in DB storage consumption.

Indexes identified as non beneficial. We had the indexes rolled back and recovered 100 GB.

Given the new pricing structure ie $40 per month for 1GB DB storage, I believe that Microsoft need to monitor each Service Maintenance for spikes in DB size.

I have raised this in the Microsoft ticket they respond
"Unfortunately there is no monitoring in place capable of keeping track of this type of increases."

I recommend that Microsoft monitor the DB Size of each instance prior to Service Maintenances and after flagging spikes for investigation.