[GitHub] [airflow] potiuk commented on issue #25372: CPU utilization 100 % loads in one worker nodes

[GitBox <gi...@apache.org>]

potiuk commented on issue #25372:
URL: https://github.com/apache/airflow/issues/25372#issuecomment-1210555788

   > hi,
   > 
   > I'm using airflow 2.3.2 in docker not Kubernetes , the worker nodes are consuming almost CPUs too, as per attachment, not sure what's doing!
   > 
   > VMs: 4core 16GB nodes: 2
   > 
   > any ideas for this? Many thanks! <img alt="810" width="555" src="https://user-images.githubusercontent.com/1091981/183844947-557cd362-4976-4940-af73-2bc655996e5d.png">
   
   I thin the question is what kind of tasks your worker are running - but only you can tell it. Possibly you spin up some processes that simply consume a lot of CPU - you need to investigate further to see what your python processes are doing. This is not a widespread problem so likely this is the code you run as part of your taskt that are doing it. 
   
   You can likely use 'trace' built in module or I actually recommend pyflame https://pyflame.readthedocs.io/en/latest/usage.html to produce interactive flamechart. See the instructions on how to run it - but it can give answer what actually takes the time.
   
   I woudl be very curious to see results of the investigation you do and flamechart. 


-- 
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.

To unsubscribe, e-mail: commits-unsubscribe@airflow.apache.org

For queries about this service, please contact Infrastructure at:
users@infra.apache.org