* Change eviction logic in NodeController and make it Zone-aware ([#28897](https://github.com/kubernetes/kubernetes/pull/28897), [@gmarek](https://github.com/gmarek))
    * Change eviction policies in NodeController:
      *  add a "partialDisruption" mode, when more than 33% of Nodes in the zone are not Ready

    int maxSize = 10;
    LocalCache<Object, Object> map =
        makeLocalCache(createCacheBuilder().concurrencyLevel(1).maximumSize(maxSize));

    // manually add elements to avoid eviction
    int originalCount = 1024;
    LinkedHashMap<Object, Object> originalMap = new LinkedHashMap<>();
    for (int i = 0; i < originalCount; i++) {
      Object key = new Object();
      Object value = new Object();

    int maxSize = 10;
    LocalCache<Object, Object> map =
        makeLocalCache(createCacheBuilder().concurrencyLevel(1).maximumSize(maxSize));

    // manually add elements to avoid eviction
    int originalCount = 1024;
    LinkedHashMap<Object, Object> originalMap = new LinkedHashMap<>();
    for (int i = 0; i < originalCount; i++) {
      Object key = new Object();
      Object value = new Object();

 *
 * <p><i>Performance notes:</i>
 *
 * <ul>
 *   <li>If you only access one end of the queue, and do use a maximum size, this class will perform
 *       significantly worse than a {@code PriorityQueue} with manual eviction above the maximum
 *       size. In many cases {@link Ordering#leastOf} may work for your use case with significantly
 *       improved (and asymptotically superior) performance.

* Kubectl
  * Display line number on JSON errors
  * Add flag -t as shorthand for --tty
* Resources
  * Improved node stability by *optionally* evicting pods upon memory pressure - [Design Doc](https://github.com/kubernetes/kubernetes/blob/release-1.3/docs/proposals/kubelet-eviction.md)
  * **alpha**: NVIDIA GPU support ([#24836](https://github.com/kubernetes/kubernetes/pull/24836), [@therc](https://github.com/therc))

    // list, it cannot be removed from it. Hence, foo is here.
    assertThat(hpackReader!!.getAndResetHeaderList()).isEqualTo(headerBlock)

    // Simulate receiving a small dynamic table size update, that implies eviction.
    bytesIn.writeByte(0x3F) // Dynamic table size update (size = 55).
    bytesIn.writeByte(0x18)
    hpackReader!!.readHeaders()
    assertThat(hpackReader!!.headerCount).isEqualTo(1)
  }

- Pod eviction now honors graceful deletion by default if no delete options are provided in the eviction request. ([#72730](https://github.com/kubernetes/kubernetes/pull/72730), [@liggitt](https://github.com/liggitt))

### Auth

    `--enforce-node-allocatable=””`
  * Hard Eviction Thresholds will be subtracted from Capacity while calculating
    Node Allocatable. This will result in a reduction of schedulable capacity in
    clusters post upgrade where kubelet hard eviction has been turned on for
    memory. To opt-out set `--experimental-allocatable-ignore-eviction=true`.
  * More details on these feature here:

- Added the correct error when eviction is blocked due to the failSafe mechanism of the `DisruptionController`. ([#133097](https://github.com/kubernetes/kubernetes/pull/133097), [@kei01234kei](https://github.com/kei01234kei)) [SIG Apps and Node]

- Fixed two regressions introduced by the `PodDisruptionConditions` feature (on by default in 1.26):
  - pod eviction API calls returned spurious precondition errors and required a second evict API call to succeed