- A pod that the Kubelet rejects was still considered as being accepted for a brief period of time after rejection, which might cause some pods to be rejected briefly that could fit on the node.  A pod that is still terminating (but has status indicating it has failed) may also still be consuming resources and so should also be considered. ([#104817](https:/...

        old API will be deprecated in a future release. See
        https://www.tensorflow.org/xla/custom_call for details.
    *   XLA:GPU reductions are deterministic by default (reductions within
        `jit_compile=True` are now deterministic).
    *   XLA:GPU works with Horovod (OSS contribution by Trent Lo from NVidia)

- Client-go: Optimized leaders renewing leases by updating leader lock optimistically without getting the record from the API server first. Also, a new metric `leader_election_slowpath_total` was added to allow users to monitor how many leader elections are updated non-optimistically.
   ([#122069](https://github.com/kubernetes/kubernetes/pull/122069), [@linxiulei](https://github.com/linxiulei))
- Locked the GA feature-gate `ConsistentHTTPGetHandlers` to default.

panic(err) } return priv, nil } // randomPoint returns a random scalar and the corresponding point using a // procedure equivalent to FIPS 186-5, Appendix A.2.2 (ECDSA Key Pair Generation // by Rejection Sampling) and to Appendix A.3.2 (Per-Message Secret Number // Generation of Private Keys by Rejection Sampling) or Appendix A.3.3 // (Per-Message Secret Number Generation for Deterministic ECDSA) followed by // Step 5 of Section 6.4.1. func randomPoint[P Point[P]](c *Curve[P], generate func([]byte)...