This post is part of Troubleshooting SQLAzure series.You can view all the posts by clicking here
SQLAzure is database as service.It also differs from our normal onpremises server in a number of ways, most important part being DTU’s.so for the rest of this post,we will focus on DTU’s.
When we procure our onpremises server,we will purchase hardware(RAM,CPU,HardDisk(Raid 1 or 5..)) based on many factors. When we procure SQLAzure database ,our database will be assigned DTU’s based on our service tier..
DTU simply put is the quantity of CPU,IO,Memory assigned to our database tier based on our subscription.It is difficult to get
actual assigned values like (1 GB ram,1 core ,raid 1 disk) but few people were able to get some rough estimates on how this DTU’s map to actual Hardware
so when procuring a database ,you will need to look at number of DTU’s you have ,as this directly affects the performance.
Below is a quote from Microsoft on the same (Emphasis in bold, mine)
When your workload exceeds the amount of any of these resources, your throughput is throttled - resulting in slower performance and timeouts.
DTU’s are directly tied to Performance,Less DTU’s equates to less performance.tFurther Microsoft says
Doubling the DTUs by increasing the performance level of a database equates to doubling the set of resource available to that database.
If you need to understand more on DTU,i recommend reading What the heck is a DTU?
Below screenshot shows rough estimates of DTU to service tier taken from Andy Mallon post
I am not sure why Microsoft reinvented the wheel and introduced a new concept called DTU,But i strongly wish Microsoft changes this and provide a chart like the one we saw above
In the next post ,we will look at tools we have to troubleshoot SQLAzure