Can you improve on the baseline (service level agreement) SLA of 99.99 % for Azure SQL Database? How can you get 99.995% as shown below, from official Microsoft documentation?
(source: https://azure.microsoft.com/en-gb/blog/understanding-and-leveraging-azure-sql-database-sla/)
This 99.99% means downtime of 52.56 minutes per year. If you opt to use Business Critical tier (or premium DTU model) with Zone redundancy configured then this improves the SLA to 99.995% – downtime cut to 26 minutes.
How and why is this possible? Well, firstly the how. Navigate to your database and go to the configure tab (here we will scale up to Business Critical).
Scroll down the page and you will see couple of extra options, specifically about read scale and more importantly the option to enable database zone redundancy.
That is it. Why is this possible. Well you can only do this within the Business Critical tier (and premium) because it does have a different architecture to that of General Purpose (or standard). The business Critical tier high availability is based on Availability Groups.
You see those secondary replicas? With the database redundant zone option we enabled we said please deploy across three different availability zones (AZ) within a region. An AZ is basically a separate data center within a given region.
Therefore it looks like the below. (You can see the separation but notice the use of Traffic Manager which is cool).
What is the draw back? Well believe it or not setting this up has no extra cost you just need to be aware that you might see some increased latency due to the distances involved, so keep an eye on it.