- The scheduler retries gated Pods more appropriately, giving them a backoff penalty too. ([#126029](https://github.com/kubernetes/kubernetes/pull/126029), [@sanposhiho](https://github.com/sanposhiho)) [SIG Scheduling]

- When the `RetryGenerateName` feature gate is enabled on the kube-apiserver,
  create requests using generateName are retried automatically by the apiserver when the generated name conflicts with an existing resource name, up to a max limit of 7 retries.
  This feature is in alpha. ([#122887](https://github.com/kubernetes/kubernetes/pull/122887), [@jpbetz](https://github.com/jpbetz))

- Improved misleading message, in case of no metrics received for the `HPA` controlled pods. ([#114740](https://github.com/kubernetes/kubernetes/pull/114740), [@kushagra98](https://github.com/kushagra98))
- Introduced new metrics removing the redundant subsystem in kube-apiserver pod logs metrics and deprecate the original ones:

has been deprecated as part of the deprecation of Heapster. Instead, you may use the [metrics server](https://github.com/kubernetes-incubator/metrics-server). It's a simplified heapster that is able to gather and serve current metrics values. It provides the Metrics API that is used by `kubectl top`, and horizontal pod autoscaler. Note that it doesn't include some features of Heapster, such as short term metrics for graphs in kube-dashboard and dedicated push sinks, which proved hard to maintain and...

## Urgent Upgrade Notes 

### (No, really, you MUST read this before you upgrade)

- For metrics `apiserver_cache_list_fetched_objects_total`, `apiserver_cache_list_returned_objects_total`, `apiserver_cache_list_total` replace `resource_prefix` label with API `group` and `resource` labels.

* Migrate the metrics/v1alpha1 API to metrics/v1beta1.  The HorizontalPodAutoscaler ([#51653](https://github.com/kubernetes/kubernetes/pull/51653), [@DirectXMan12](https://github.com/DirectXMan12))
    * controller REST client now uses that version.  For v1beta1, the API is now known as
    * resource-metrics.metrics.k8s.io.

- `scheduler`: handling of unschedulable pods because a `ResourceClass` is missing
  is a bit more efficient and no longer relies on periodic retries. ([#120213](https://github.com/kubernetes/kubernetes/pull/120213), [@pohly](https://github.com/pohly))

## Dependencies

### Added
- cloud.google.com/go/dataproc/v2: v2.0.1

- The nftables kube-proxy mode now has its own metrics rather than reporting
  metrics with "iptables" in their names. ([#124557](https://github.com/kubernetes/kubernetes/pull/124557), [@danwinship](https://github.com/danwinship)) [SIG Network and Windows]

    - [(No, really, you MUST read this before you upgrade)](#no-really-you-must-read-this-before-you-upgrade)
  - [Deprecations](#deprecations)
  - [Removed and deprecated metrics](#removed-and-deprecated-metrics)
    - [Removed metrics](#removed-metrics)
    - [Deprecated metrics](#deprecated-metrics)
  - [Notable Features](#notable-features)
  - [API Changes](#api-changes)
  - [Detailed Bug Fixes And Changes](#detailed-bug-fixes-and-changes)

* The Metrics Server now exposes metrics via the /metric endpoint. These metrics are in the prometheus format.  ([#57456](https://github.com/kubernetes/kubernetes/pull/57456), [@kawych](https://github.com/kawych))