Common DNS name for Azure SQL
SQL AzureDB is built apon Always On Availability Groups from what I have seen in SQL Server Presentations. Part of the SQL Server Availability groups there is a Common DNS name which sits over the top of all the SQL Databases in the Availability group, you connect to this "Common DNS Name" so in the event of a failover, the client connections get automatically redirected to the correct server/db.
SQL AzureDB does not present this as an option to connect to via the Azure portal, which is why I was talking to the GM for Azure Cat team, which he agreed it *should* be something Microsoft might be able to offer to its customers as they would only need to expose the listener.
Again, we have SQL AzureDB in production, what I am trying to achieve is a *common* name to set the client to connect to and on failover there is no connection string change. We got from support there is a similar capacity of "common DNS name" on Azure SQL is going to preview, please share us more details.
Ayo Olubeko commented
This is now possible by using the failover group capabilities of Azure SQL Server: https://azure.microsoft.com/en-us/blog/azure-sql-database-now-supports-transparent-geographic-failover-of-multiple-databases-featuring-automatic-activation/
Adam Weigert commented
I'd also like a common name that moves with the primary database, and support for read/write-intent. This would make the applications zero-touch during a database server maneuver such as shifting a region or changing the server it is associated with. I'd take a simple trafficmanager profile at this point, having to modify the application connection strings while automation is possible, it can be highly prone to error especially since we use source control config of the app settings / connection strings which can't be easily automatically updated so the new release pushes the right configuration.
Oleg Ignat commented
SQL Azure is not built on top of traditional Always On Availability Groups but a somewhat similar technology. It already provides physical SQL instance abstraction in case a particular node in the cluster experiences an outage. Namely you continue to be connected to <your server name>.database.windows.net, regardless which cluster or data center in a region this name is pointing to at any given moment.
I'd love to hear more about this ask. Please drop me an email at email@example.com.