Traffic Manager for SQL Database
The traffic manager is a great way to account for failover between web nodes in different data centers. However, an analogous paradigm doesn't exist for databases. In the premium tier, you can enable active geo-replication. However, all applications dependent on the primary either need to have their connection strings changed over when failing over between data centers, or they have to have some sophisticated retry logic built in at the application tier. It would be great if applications could point to a single, consistent endpoint and the geo-failover could be handled at a higher level the way it is with the Traffic Manager.
This is not currently planned at this time. Will keep this item open for voting and comments.
Jacob Golden commented
Was this not addressed with the Failover Groups functionality at the SQL Server level in Azure SQL?
This would be a very powerful & useful feature. Traffic Manager folks, get this done please, the sooner the better.
Any update on this ask?
Luke Ballantine commented
I completely agree.
I was quite disappointed to find that ATM lets you select a primary and secondary web app, but doesn't give you the option to select a primary and secondary SQL Azure DB. It seems like the dev team really missed a trick there. I would have been a super-simple geo-redundant DR solution. As it stands now, you've got to roll-your-own monitor app for DB failover that kills the SQL replication so your secondary DB becomes read-write.