- Overall, enable the feature of `PreferNominatedNode` will  improve the performance of scheduling where preemption might frequently happen, but in theory, enable the feature of `PreferNominatedNode`, the pod might not be scheduled to the best candidate node...

- The new `dra_resource_claims_in_use` kubelet metrics reported active `ResourceClaims`, overall and by driver. ([#131641](https://github.com/kubernetes/kubernetes/pull/131641), [@pohly](https://github.com/pohly)) [SIG Architecture, Instrumentation, Node and Testing]

   ([#123271](https://github.com/kubernetes/kubernetes/pull/123271), [@neolit123](https://github.com/neolit123))
- kubeadm: improved the overall logic, error handling, and output messages when waiting for the kubelet and API server `/healthz` endpoints to return `OK`. The kubelet and API server checks no longer ran in parallel, but one after another (in serial).

  - If all QueueingHintFn return QueueSkip, it puts this pod back to the unschedulable pod pool
  
  Having appropriate QueueingHintFn contributes to reducing useless retries and thus improves the overall scheduler's performance.
  
  **How can I migrate?**
  
  For backward compatibility, nil `QueueingHintFn` is treated as always returning QueueAfterBackoff.