3

At this moment we are using a serverless python function app to run multiple azure functions. When I run a certain function locally (VScode), the execution-time is usually around a minute.

However, in the deployed version the function always reaches the 10-minute time-out and stops running. The function-app is currently using a consumption plan and I have allocated the maximum amount of FUNCTIONS_WORKER_PROCESS_COUNT. Regardless, I am still having the same performance issues.

Another thing I noticed that the function does not allows parallel runs. It seems to wait for one call to finish before running another one. From my understanding this should not be the case as the consumption plan should be able to run concurrently and allocate the resources it needs automatically.

BrahimEG
  • 41
  • 1
  • What is the trigger? You are right that the consumption plan allows parallel runs, but some workloads require some configuration to maximize perf. Happy to give you some recommendations if you can describe what the function does and what the trigger is. – Anthony Chu Jul 30 '20 at 00:27
  • It is an Http-triggered function which runs an ML script that uses multiple libraries. On all my local devices one run usually takes about a minute. Thanks for the reply! PS: I already read up about the concept of a cold-start and that does not seem to be the problem in my case. – BrahimEG Jul 30 '20 at 07:38

0 Answers0