- Fixes a problem with 63-second or 1-second connection delays with some VXLAN-based
  network plugins which was first widely noticed in 1.16 (though some users saw it
  earlier than that, possibly only with specific network plugins). If you were previously
  using ethtool to disable checksum offload on your primary network interface, you should

### Action Required

* [action required] Default Flexvolume plugin directory for COS images on GCE is changed to `/home/kubernetes/flexvolume`. ([#58171](https://github.com/kubernetes/kubernetes/pull/58171), [@verult](https://github.com/verult))

Africa/Dakar Africa/Dar_es_Salaam Africa/Djibouti Africa/Douala Africa/El_Aaiun Africa/Freetown Africa/Gaborone Africa/Harare Africa/Johannesburg Africa/Juba Africa/Kampala Africa/Khartoum Africa/Kigali Africa/Kinshasa Africa/Lagos Africa/Libreville Africa/Lome Africa/Luanda Africa/Lubumbashi Africa/Lusaka Africa/Malabo Africa/Maputo Africa/Maseru Africa/Mbabane Africa/Mogadishu Africa/Monrovia Africa/Nairobi Africa/Ndjamena Africa/Niamey Africa/Nouakchott Africa/Ouagadougou Africa/Porto-Novo Africa/Sao_Tome...

- Sometimes, the scheduler incorrectly placed a pod in the "unschedulable" queue instead of the "backoff" queue. This happened when some plugin previously declared the pod as "unschedulable" and then in a later attempt encounters some other error. Scheduling of that pod then got delayed by up to five minutes, after which periodic flushing moved the pod back into the "active" queue. ([#120334](https://github.com/kubernetes/kubernetes/pull/120334),...

* set EnableHTTPSTrafficOnly in azure storage account creation ([#64957](https://github.com/kubernetes/kubernetes/pull/64957), [@andyzhangx](https://github.com/andyzhangx))
* Re-adds `pkg/generated/bindata.go` to the repository to allow some parts of k8s.io/kubernetes to be go-vendorable. ([#65985](https://github.com/kubernetes/kubernetes/pull/65985), [@ixdy](https://github.com/ixdy))

Before updating to 1.9, you are strongly recommended to back up your etcd data. Consult the installation procedure you are using (kargo, kops, kube-up, kube-aws, kubeadm etc) for specific advice.

* add GRS, RAGRS storage account type support for azure disk ([#55931](https://github.com/kubernetes/kubernetes/pull/55931), [@andyzhangx](https://github.com/andyzhangx))
* Fix a typo in prometheus-to-sd configuration, that drops some stackdriver metrics. ([#56473](https://github.com/kubernetes/kubernetes/pull/56473), [@loburm](https://github.com/loburm))

  - [Changelog since v1.20.5](#changelog-since-v1205)
  - [Important Security Information](#important-security-information-1)
    - [CVE-2021-25735: Validating Admission Webhook does not observe some previous fields](#cve-2021-25735-validating-admission-webhook-does-not-observe-some-previous-fields)
  - [Changes by Kind](#changes-by-kind-9)
    - [API Change](#api-change-2)
    - [Feature](#feature-5)
    - [Bug or Regression](#bug-or-regression-9)

In some cases, for pure data validation and processing, you can get performance improvements of **20x** or more. This means 2,000% or more. 🤯

When you use **FastAPI**, there's a lot more going on, processing the request and response, handling dependencies, executing **your own code**, and particularly, **waiting for the network**. But you will probably still get some nice performance improvements just from the upgrade.

  // The set of all endpoints is the union of all subsets. Addresses are placed into
  // subsets according to the IPs they share. A single address with multiple ports,
  // some of which are ready and some of which are not (because they come from
  // different containers) will result in the address being displayed in different
  // subsets for the different ports. No address will appear in both Addresses and