- Rename PluginContext to CycleState in the scheduling framework ([#83430](https://github.com/kubernetes/kubernetes/pull/83430), [@draveness](https://github.com/draveness))
- Some scheduler extender API fields are moved from `pkg/scheduler/api` to `pkg/scheduler/apis/extender/v1`. ([#83262](https://github.com/kubernetes/kubernetes/pull/83262), [@Huang-Wei](https://github.com/Huang-Wei))

* Kubelet will now set extended resource capacity to zero after it restarts. If the extended resource is exported by a device plugin, its capacity will change to a valid value after the device plugin re-connects with the Kubelet. If the extended resource is exported by an external component through direct node status capacity patching, the component should repatch the field after kubelet...

- Previously, the scheduling queue didn't notice any extenders' failures, potentially resulting in missed cluster events and Pods rejected by Extenders being stuck in the unschedulable pod pool for up to 5 minutes in the worst-case scenario. Now, the scheduling queue notices extenders' failures and requeues Pods rejected by Extenders appropriately.

      - [Load Balancing](#load-balancing)
    - [**Node Components**](#node-components)
      - [Container Runtime Interface](#container-runtime-interface)
    - [**Scheduling**](#scheduling-1)
      - [Scheduler Extender](#scheduler-extender)
    - [**Storage**](#storage-1)
      - [Local Storage](#local-storage)
      - [Volume Plugins](#volume-plugins)
      - [Metrics](#metrics)
    - [**Other notable changes**](#other-notable-changes-16)

- If scheduler extender filtered a not found node, current scheduling round for this pod will just be skipped.
  ([#79641](https://github.com/kubernetes/kubernetes/pull/79641), [@yqwang-ms](https://github.com/yqwang-ms))

- Print NotReady when pod is not ready based on its conditions. ([#88240](https://github.com/kubernetes/kubernetes/pull/88240), [@soltysh](https://github.com/soltysh)) [SIG CLI]
- Scheduler Extender API is now located under k8s.io/kube-scheduler/extender ([#88540](https://github.com/kubernetes/kubernetes/pull/88540), [@damemi](https://github.com/damemi)) [SIG Release, Scheduling and Testing]

- Use sync.map to scale ecache better ([#66862](https://github.com/kubernetes/kubernetes/pull/66862), [@resouer](https://github.com/resouer))
- Extender preemption should respect IsInterested() ([#66291](https://github.com/kubernetes/kubernetes/pull/66291), [@resouer](https://github.com/resouer))

* Added support to minimize sending verbose node information to scheduler extender by sending only node names and expecting extenders to cache the rest of the node information ([#41119](https://github.com/kubernetes/kubernetes/pull/41119), [@sarat-k](https://github.com/sarat-k))

    * Reduce logging agent buffer chunk limit to 512k to optimize performance.

  - By default, injected tokens are given an extended lifetime so they remain valid even after a new refreshed token is provided. The metric `serviceaccount_stale_tokens_total` can be used to monitor for workloads that are depending on the extended lifetime and are continuing to use tokens even after a refreshed token is provided to the container. If that metric indicates no existing workloads are depending on extended lifetimes, injected token lifetime can be shortened to 1 hour...