Microsoft takes data integrity very seriously. While there are some traditional techniques used by DBAs in SQL Server to monitor data integrity (e.g. DBCC CHECKDB) and various methods to recover from database corruptions, the Azure SQL Database engineering team has been able to develop new techniques that can handle some classes of corruption automatically and without data loss. These techniques are used as part of the service to avoid data loss and downtime in the cases where they can be avoided.
This blog post outlines some of those techniques, how they work, and what impact it has on customers concerned with what steps they should take to safeguard their data in Azure SQL Database.
Read about it in the Azure blog.