// (i.e. it's a no-op). A null preferred scheduling term matches no objects (i.e. is also a no-op).
message PreferredSchedulingTerm {
  // Weight associated with matching the corresponding nodeSelectorTerm, in the range 1-100.
  optional int32 weight = 1;

  // A node selector term, associated with the corresponding weight.
  optional NodeSelectorTerm preference = 2;
}

- Kube-proxy service health returns http header `X-Load-Balancing-Endpoint-Weight` with number of local endpoints. The same information is still available in response body JSON `payload.LocalEndpoints`. ([#118999](https://github.com/kubernetes/kubernetes/pull/118999), [@cezarygerard](https://github.com/cezarygerard))

- Introduce `v1beta3` API for scheduler. This version 
  - increases the weight of user specifiable priorities.
  The weights of following priority plugins are increased
    - `TaintTolerations` to 3 - as leveraging node tainting to group nodes in the cluster is becoming a widely-adopted practice

- The deprecated flag `--hard-pod-affinity-symmetric-weight` and `--scheduler-name` have been removed from `kube-scheduler`. Use `ComponentConfig` instead to configure those parameters. ([#102805](https://github.com/kubernetes/kubernetes/pull/102805), [@ahg-g](https://github.com/ahg-g))

  Previously, for IPv6, if the podSubnet had a mask lower than /112, kubeadm calculated a node-mask to be multiple of eight and splitting the available bits to maximise the number used for nodes. ([#95723](https://github.com/kubernetes/kubernetes/pull/95723), [@aojea](https://github.com/aojea)) [SIG Cluster Lifecycle]