Praveen Krishna comments

Results 23 comments of


                                            Praveen Krishna

[FG:InPlacePodVerticalScaling] Performance degradation in latency-sensitive services due to CPU affinity loss upon guaranteed QoS Pod scaling down

/cc @pravk03

[FG:InPlacePodVerticalScaling] Performance degradation in latency-sensitive services due to CPU affinity loss upon guaranteed QoS Pod scaling down

The current [proposal](https://docs.google.com/document/d/19-yzI41L6_XRj6l_27ylWSc114fzgP301FhxCPf0Dbg/edit?tab=t.0) and [POC](https://github.com/kubernetes/kubernetes/pull/129719) by @esotsal is based on this [previous SIG Node discussion](https://www.youtube.com/watch?v=RuqzXH3liqg) where it was discussed that guaranteeing initial CPU allocation (promised CPUs) for containers is sufficient....

Implicit tolerations

/cc > I see the usecase for the devices or non-core resources. I'm thinking especially about exclusive CPUs, but memory/hugepages can have the same problem. @ffromani Considering that we have...