ABHINAV INAMDAR

Hi @natasha41575 , I'd like to work on this. I see that the resize logic is handled in doPodResizeAction (in pkg/kubelet/kuberuntime/kuberuntime_container.go). I can work on removing the best-effort memory check...

Hi @natasha41575, Regarding the approach to 'move the check into the runtime, I see two potential paths: 1. Minimize the Race: Move the memory validation in Kubelet to immediately before...

I understand that a KEP is required to formalize the CRI changes and coordinate with containerd/CRI-O maintainers. My initial technical plan involves: **Atomic Validation**: Standardizing a requirement in CRI for...

I agree that phased shrinking is a valuable strategy for inducing memory reclaim gracefully. It would definitely improve the safety profile, especially on non-Linux platforms where atomic checks might not...

Thanks @lmktfy. I will ensure the alternatives section of the KEP explicitly details the phased shrinking approach and the trade-offs that led us to prioritize the atomic runtime check. I'll...

Hi @natasha41575, You're right, I am new to the Kubernetes org, and I respect your decision to prioritize stability on such a complex issue. I definitely don't want to become...

Hi @danwinship, I’ve been working on the pod reduction strategy you suggested. I have a local implementation ready that optimizes the NewModel topology: 1. Change: Refactored NewModel to support a...

Hi @sivamalla42, I can help dig into this. Could you please provide a few more details to help diagnose the crash? 1. Exact Versions 2. Pod Details & Events Looking...