- Added new option to the `InterPodAffinity` scheduler plugin to ignore existing
  pods` preferred inter-pod affinities if the incoming pod has no preferred inter-pod
  affinities. This option can be used as an optimization for higher scheduling throughput
  (at the cost of an occasional pod being scheduled non-optimally/violating existing
  pods preferred inter-pod affinities). To enable this scheduler option, set the

- User can now use component config to configure NodeLabel plugin for the scheduler framework. ([#84297](https://github.com/kubernetes/kubernetes/pull/84297), [@liu-cong](https://github.com/liu-cong))
- Optimize inter-pod affinity preferredDuringSchedulingIgnoredDuringExecution type, up to 4x in some cases. ([#84264](https://github.com/kubernetes/kubernetes/pull/84264), [@ahg-g](https://github.com/ahg-g))

- The Kubernetes API server merged selectors built from `matchLabelKeys` into the `labelSelector` of `topologySpreadConstraints`, aligning Pod Topology Spread behavior with Inter-Pod Affinity. To prevent breaking existing Pods using `matchLabelKeys`, this scheduler behavior was preserved until v1.34. Upgrades from v1.32 to v1.34 should be done incrementally (v1.32 → v1.33 → v1.34), ensuring Pods created at v1.32 with `matchLabelKeys`...

  * [Nodes](https://kubernetes.io/docs/concepts/configuration/assign-pod-node/#node-affinity-beta-feature) and [pods](https://kubernetes.io/docs/concepts/configuration/assign-pod-node/#inter-pod-affinity-and-anti-affinity-beta-feature) now support affinity and anti-affinity

- Add improved load balancer naming.

### SIG-scheduling

SIG Scheduling development efforts have been primarily focused on improving performance and reliability of the scheduler.
- Performance of the inter-pod affinity/anti-affinity feature is improved over 100X via algorithmic optimization.

<p>
Like arrays, slices are always one-dimensional but may be composed to construct
higher-dimensional objects.
With arrays of arrays, the inner arrays are, by construction, always the same length;
however with slices of slices (or arrays of slices), the inner lengths may vary dynamically.
Moreover, the inner slices must be initialized individually.
</p>

<h3 id="Struct_types">Struct types</h3>

<p>

- github.com/go-kit/log: [v0.1.0](https://github.com/go-kit/log/tree/v0.1.0)
- github.com/gofrs/uuid: [v4.0.0+incompatible](https://github.com/gofrs/uuid/tree/v4.0.0)
- github.com/josharian/intern: [v1.0.0](https://github.com/josharian/intern/tree/v1.0.0)
- github.com/jpillora/backoff: [v1.0.0](https://github.com/jpillora/backoff/tree/v1.0.0)
- github.com/nxadm/tail: [v1.4.4](https://github.com/nxadm/tail/tree/v1.4.4)

* Timeout docker queries to prevent node going NotReady indefinitely. ([#56967](https://github.com/kubernetes/kubernetes/pull/56967), [@jsravn](https://github.com/jsravn))
* Fix bug allowing garbage collector to enter a broken state that could only be fixed by restarting the controller-manager. ([#61201](https://github.com/kubernetes/kubernetes/pull/61201), [@jennybuckley](https://github.com/jennybuckley))

- The `Service.spec.ipFamilyPolicy` field is now *required* in order to create or update a Service as dual-stack.  This is a breaking change from the beta behavior.  Previously the server would try to infer the value of that field from either `ipFamilies` or `clusterIPs`, but that caused ambiguity on updates.  Users who want a dual-stack Service MUST specify `ipFamilyPolicy` as either "PreferDualStack" or "RequireDualStack". ([#96684](https://github....

* Fixes [#49418](https://github.com/kubernetes/kubernetes/pull/49418) where kube-controller-manager can panic on volume.CanSupport methods and enter a crash loop. ([#49420](https://github.com/kubernetes/kubernetes/pull/49420), [@gnufied](https://github.com/gnufied))