0

I've a kubernetes deployments on GKE. I've set CPU request and limit to be the same ~700m. When i observe the performances of the pods consuming a single message queue (RabbitMQ), can confirm the messages have very little deviation between themselves, each pod consistently performs differently.

Graph 1

The graph shows the time taken to process messages by each pod excluding network delays, ie. this is total time only taken to process the message from the point it was received by pod consumer.

Can confirm we allocated enough compute resources. Still we are seeing consistent differences in time taken to process messages. What could be the reason for this sort of observation on Kubernetes. Thank you.

P.S. : The deployment show here runs a python code.

1 Answer 1

1

This looks like something that needs deeper investigation in your GKE cluster. I suggest you to raise a Public Issue Tracker with your project number (just numbers). Be sure not to write your project ID instead. Once created, provide me the link to the issue.

1

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .