- Fixes a regression exposed in 1.27 by kubectl switching to openapi v3 by resolving a flake in openapi v3 aggregation ([#119839](https://github.com/kubernetes/kubernetes/pull/119839), [@atiratree](https://github.com/atiratree)) [SIG API Machinery]

* Fix regression in the CRI: do not add a default hostname on short image names ([#58955](https://github.com/kubernetes/kubernetes/pull/58955), [@runcom](https://github.com/runcom))

* use containing API group when resolving shortname from discovery ([#58741](https://github.com/kubernetes/kubernetes/pull/58741), [@dixudx](https://github.com/dixudx))

* Bug fixes

  * Changed Kubelet default image-gc-high-threshold to 85% to resolve a conflict with default settings in docker that prevented image garbage collection from resolving low disk space situations when using devicemapper storage. ([#40432](https://github.com/kubernetes/kubernetes/pull/40432), [@sjenning](https://github.com/sjenning))

* Add 'X-Content-Type-Options: nosniff" to some error messages ([#37190](https://github.com/kubernetes/kubernetes/pull/37190), [@brendandburns](https://github.com/brendandburns))
* Fixes bug in resolving client-requested API versions ([#38533](https://github.com/kubernetes/kubernetes/pull/38533), [@DirectXMan12](https://github.com/DirectXMan12))

*   `tf.debugging`
    *   Add `tf.debugging.enable_check_numerics()` and
        `tf.debugging.disable_check_numerics()` to help debugging the root
        causes of issues involving infinities and `NaN`s.
*   `tf.distribute`
    *   Custom training loop support on TPUs and TPU pods is available through
        `strategy.experimental_distribute_dataset`,

- Fixed a bug in `kubectl` that caused the wrong result length when using `--chunk-size` and `--selector` together. ([#110652](https://github.com/kubernetes/kubernetes/pull/110652), [@Abirdcfly](https://github.com/Abirdcfly))
- Fixed a bug involving Services of type `LoadBalancer` with multiple IPs and a `LoadBalancerSourceRanges` that overlaps the node IP. ([#109826](https://github.com/kubernetes/kubernetes/pull/109826), [@danwinship](https://github.com/danwinship))

- Fixed node data lost in kube-scheduler for clusters with imbalance on number of nodes across zones ([#93355](https://github.com/kubernetes/kubernetes/pull/93355), [@maelk](https://github.com/maelk))
- Fixed several bugs involving the IPFamily field when creating or updating services
  in clusters with the IPv6DualStack feature gate enabled.
  
  Beware that the behavior of the IPFamily field is strange and inconsistent and will

* advanced audit: shutdown batching audit webhook gracefully ([#50577](https://github.com/kubernetes/kubernetes/pull/50577), [@crassirostris](https://github.com/crassirostris))
* Add Priority admission controller for monitoring and resolving PriorityClasses. ([#49322](https://github.com/kubernetes/kubernetes/pull/49322), [@bsalamat](https://github.com/bsalamat))

- kube- and cloud-controller-manager can now listen on ports up to 65535 rather than 32768, solving problems with operating systems that request these higher ports.. ([#65860](https://github.com/kubernetes/kubernetes/pull/65860), [@sttts](https://github.com/sttts))

## Changelog since v1.26.0

## Changes by Kind

### API Change

- The list-type of the alpha resourceClaims field introduced to Pods in 1.26.0 was modified from "set" to "map", resolving an incompatibility with use of this schema in CustomResourceDefinitions and with server-side apply. ([#114617](https://github.com/kubernetes/kubernetes/pull/114617), [@JoelSpeed](https://github.com/JoelSpeed)) [SIG API Machinery]