We're hosting a site on Azure as an App Service and today we had an unexpected memory behavior.

We had about 200 concurrent users (which for our application means about 3K requests/min) and it was running fine until all of a sudden the memory usage peaked. It went high for a short while but then quickly went back to previous values.

This also caused our CPU to hit 97% for a short time (a few seconds) which caused the response time to rise and also some HTTP-500-responses were sent.

We are running on one S1 instance, and we are definitely prepared to scale up/out if necessary, but we'd still like to understand more about this peak. Does anyone recognize this behavior?

I could also add that our normal memory-usage is around 60-70%, is this high? It doesn't go down even during the night when we have no users.

Additional info: Our DTU-percentage went from 25% to 50%. This is probably related. Is there a functionality to load an entire DB-table into memory if it is I/O-intensive for a short while?

enter image description here

  • We're not going to be able to tell you what occured without more details on what your app is doing. Have you considered hooking your app up to AppInsights so you can collect data on what is being run during these spikes? – Sam Cogan Nov 21 '17 at 17:11

Your Answer

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

Browse other questions tagged or ask your own question.