cancel
Showing results for 
Show  only  | Search instead for 
Did you mean: 

Profiling CPU issues

IlCrux
Visitor

Hi,

I'm trying to profile some CPU spikes I got during the day on an Azure App Service.

Here's basically the graph of the CPU consume over a single istance:

MarcoTC_0-1706626256210.pngMarcoTC_1-1706626447417.png

If I click View detailed CPU breakdown, I got this:

MarcoTC_2-1706626482746.png

But it seems the most CPU usage is by "Background thread CPU"? Which (if I understand it correctly) related to PipelineRuntime.ProcessRequestNotification:

MarcoTC_3-1706626546987.png

Which is the entry point of the dynatrace profiler 😮

Am I correct? It seems profiler introduce the main CPU analying it? Do I need to configure out somethings different?

Thanks

 

3 REPLIES 3

IlCrux
Visitor

No one have a clue?

BUMP?

Julius_Loman
DynaMight Legend
DynaMight Legend

Is there any chance your application uses native libraries? Also, there are OneAgent features for .NET you might not have enabled (see OneAgent features enabled in your environment).

Anyway in .NET, Stack Traces are not collected very often, especially compared to Java applications. I'd recommend looking at a wider interval when analysing it with method hotspots. You have just 135 samples of stack traces in your timeframe and this is not enough for any conclusion.

Certified Dynatrace Master | Alanata a.s., Slovakia, Dynatrace Master Partner

Featured Posts