Several of our Azure VMs were automatically shutdown due to an automated Azure process where the reads/writes are to slow or do not complete in 120 seconds, inclusive of retries, the behavior of the VM is to shutdown to protect customer data and prevent corruption.
Although the Azure platform continues to monitor the storage services to determine if the service is restored so it can auto start the VM the customer is never notified this occurs. I discovered this myself when I noticed application downtime.
A great process to protect your customers, but horrible execution so your customers are aware this is happening.
Please generate a notification system so customers can provide an alerting address for these types of Microsoft hardware issues which the customer has no control over.