*   Keras-based LSTM models must be converted with an explicit batch size in
        the input layer.

## Bug Fixes and Other Changes

### TF Core:

*   Set `tf2_behavior` to 1 to enable V2 for early loading cases.
*   Add `execute_fn_for_device function` to dynamically choose the
    implementation based on underlying device placement.
*   Eager:
    *   Add `reduce_logsumexp` benchmark with experiment compile.

- Added the feature gates `StrictCostEnforcementForVAP` and `StrictCostEnforcementForWebhooks` to enforce the strict cost calculation for CEL extended libraries. It is strongly recommended to turn on the feature gates as early as possible. ([#124675](https://github.com/kubernetes/kubernetes/pull/124675), [@cici37](https://github.com/cici37)) [SIG API Machinery, Auth, Node and Testing]

The focus of this release is **compatibility** with Pydantic v1 and v2, to make sure your current apps keep working. Later there will be more focus on refactors, correctness, code improvements, and then **performance** improvements. Some third-party early beta testers that ran benchmarks on the beta releases of FastAPI reported improvements of **2x - 3x**. Which is not bad for just doing `pip install --upgrade fastapi pydantic`. This was not an official benchmark and I didn't check it myself, but it's...

- Endpoint slice controller now mirrors parent's service label to its corresponding endpoint slices....

- Fixed: deleted a non-existent Azure disk issue. ([#107406](https://github.com/kubernetes/kubernetes/pull/107406), [@andyzhangx](https://github.com/andyzhangx))
- Fixed: do not return early in the node informer when there is no change of the topology label. ([#108149](https://github.com/kubernetes/kubernetes/pull/108149), [@nilo19](https://github.com/nilo19))

### API Change

- Added the feature gates `StrictCostEnforcementForVAP` and `StrictCostEnforcementForWebhooks` to enforce the strct cost calculation for CEL extended libraries. It is strongly recommended to turn on the feature gates as early as possible. ([#124676](https://github.com/kubernetes/kubernetes/pull/124676), [@cici37](https://github.com/cici37)) [SIG API Machinery, Auth, Node and Testing]

    <magic priority="40">
      <match value="function [a-zA-Z][A-Za-z0-9_]{0,62}[\\r\\n]" type="regex" offset="0"/>
    </magic>
    <!-- Two matlab-style comments fairly early in the file -->
    <magic priority="25">
      <match value="%" type="string" offset="0">
         <match value="\n%" type="string" offset="2:120"/>
      </match>
      <match value="%" type="string" offset="0">

nodes with pods may see pods that fail due to node admission

1.22 addressed a long-standing issue in the Kubelet where terminating pods were [vulnerable to race conditions](https://github.com/kubernetes/kubernetes/pull/102344) leading to early shutdown, resource leaks, or long delays in actually completing pod shutdown. As a consequence of this change the Kubelet now correctly takes into account the resources of running and terminating pods when deciding to accept new pods, since terminating...

### Increase the Kubernetes support window to one year

As of Kubernetes 1.19, bugfix support via patch releases for a Kubernetes minor release has increased from 9 months to 1 year.

A survey conducted in early 2019 by the working group (WG) Long Term Support (LTS) showed that a significant subset of Kubernetes end-users fail to upgrade within the previous 9-month support period.