- apiserver_flowcontrol_request_wait_duration_seconds 
  - apiserver_flowcontrol_current_executing_seats
  - apiserver_flowcontrol_nominal_limit_seats
  - apiserver_flowcontrol_rejected_requests_total
  - apiserver_flowcontrol_dispatched_requests_total
  - apiserver_flowcontrol_current_inqueue_requests

* Requests with the query parameter `?watch=` are treated by the API server as a request to watch, but authorization and metrics were not correctly identifying those as watch requests, instead grouping them as list calls. ([#48583](https://github.com/kubernetes/kubernetes/pull/48583), [@smarterclayton](https://github.com/smarterclayton))

- Kube-apiserver: fixes a 1.27+ regression in watch stability by serving watch requests without a resourceVersion from the watch cache by default, as in <1.27 (disabling the change in #115096 by default). This mitigates the impact of an etcd watch bug (https://github.com/etcd-io/etcd/pull/17555). If the 1.27 change in #115096 to serve these requests from underlying storage is still desired despite the impact on watch stability, it can be re-enabled with a `Watc...

  - `priority_level_request_count_samples` is replaced with `priority_level_request_utilization`, which samples every nanosecond rather than every millisecond; the old metric conveyed utilization despite its name.
  - `priority_level_request_count_watermarks` is removed.

- When the kubelet fails to assign CPUs to a Pod because there less available CPUs than the Pod requests, the error message changed from
  `not enough cpus available to satisfy request` to `not enough cpus available to satisfy request: <num_requested> requested, only <num_available> available`. ([#121059](https://github.com/kubernetes/kubernetes/pull/121059), [@matte21](https://github.com/matte21))

- Deprecated the `apiserver_request_slo_duration_seconds` metric for v1.27 in favor of `apiserver_request_sli_duration_seconds` for naming consistency purposes with other SLI-specific metrics and to avoid any confusion between SLOs and SLIs. ([#112679](https://github.com/kubernetes/kubernetes/pull/112679), [@dgrison...

- Kubeadm: make sure that a variety of API server requests are retried during "init", "join", "upgrade", "reset" workflows. Prior to this change some API server requests, such as, creating or updating ConfigMaps were "one-shot"...