I am new to Azure architecture. I want to manage failover between two data centers.


Data center A has Web apps, storage account, SQL database.

Data center B has same services Web apps, storage account, SQL database.

When one of the service goes down from Data center A, traffic manager should send request to Data center B where all services should up and provide response to requests.


The actual implementation you choose for your app is really up to you (and the topic itself is broad and opinion-soliciting). But, from an objective perspective, there are features built in to Azure that have multi-geo support that you can take advantage of, regarding the areas you're investigating:

  • Traffic Manager. As you're using Web apps: This service offers traffic-balancing between multiple web apps, whether in the same region or different regions, and offers three modes: Failover, round robin, and performance (closest, latency-wise). More info here.
  • Azure Storage. As you're using storage: There are multiple tiers available, one being Read-access geographically-redundant (RA-GRS). Storage is asynchronously replicated to an alternate region. Should the primary region become unavailable (or if you simply want to read from secondary for performance or other reasons), you may choose to perform your storage reads from the secondary location, via a slightly-modified storage endpoint (yourstorage-secondary.blob.core.windows.net for example). Note: You cannot choose this location; Azure regions are paired.
  • Azure SQL Database. Since you're using SQL Database: Standard and Premium tiers offer standard geo-replication, while Premium tier offers Active Geo-replication with up to 4 readable secondaries. More info here.

How (or if) you choose to use these features is up to you...

  • Thanks for sharing solution. can we sync or replicate event hub? I am using event hub too in my application. – sam Mar 24 '16 at 5:08

For this Azure traffic manager is good option to failover trick. You can have them DC A and B under whatever strategy you want like failover/round robin.


I have found most suitable answer for this question here

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.