CPU Utilization over 90%

We have installed SQLWatch on several instances on azure. We have had this issue with SQLWatch consuming CPU most of the time and its always above 90%.

Currently i am trying to diagnose issue on a vm with 8 vCPU and 32gb RAM. Hardisk space is well over 100gb.

SQLWATCH version is 2.5.7460.35921

I had to shutdown the SQLWatch because it was draining the CPU. I was wondering if anyone else is facing the same issue here?

Can you please be more specific, which jobs are causing the CPU go up? When yous say “shutdown SQLWATCH?” What do you mean? SQLWATCH are essentially just agent jobs that collect some data every few minutes.
Quite possibly it is the index jobs, if you have a lot of tables and indexes.

I don’t know which jobs are causing the Highest CPU utilisation. By shutdown i meant to say I had to put the SQLWATCH database “Offline” I ran a query to check which database was consuming so much CPU and SQLWATCH was always showing high. The question is this a normal behaviour ? attached an image for reference.

There are not lots of tables or databases.


All this screenshot says is that the SQLWATCH is the most active database, and of ALL the CPU utilisation by SQL Server, it had 95%. It does not mean it caused the CPU to go up to over 95%. Can you show me the actual CPU utilisation either from the SQLWATCH itself or from the performance counter, or even from the task manager?

So i had to go back to the drawing board and look at the whole thing all over again. SQLWATCH was not an issue or causing the CPU high consumption. However, i noticed that there are two jobs agents , SQLWATCH_INTERNAL_CHECK and SQLWATCH_LOGGER_PERFORMANCE causing the cpu to spike 100% every 1 minute BUT for less than a second which i assume could be a normal behaviour.

I will close this discussion for now and thank you for your imput.

1 Like

Did you look at this topic?: https://docs.sqlwatch.io/#performance-overhead

As for the SQLWATCH_INTERNAL_CHECK and SQLWATCH_LOGGER_PERFORMANCE jobs, the performance will be further improved in version 3.0.