- A new kubelet metric `image_pull_duration_seconds` was added. The metric tracks the duration (in seconds) it takes for an image to be pulled, including the time spent in the waiting queue of image puller. The metric is broken down by bucketed image size. ([#121719](https://github.com/kubernetes/kubernetes/pull/121719), [@ruiwen-zhao](https://github.com/ruiwen-zhao))

   The `read_vs_write_current_requests` metric histogram is adjusted, in the case of using API Priority and Fairness instead of max-in-flight, to divide by the relevant limit: sum of queue capacities for waiting requests, sum of seat limits for executing requests. ([#110164](https://github.com/kubernetes/k...

  
  ### Exoscale
  - Add support for Exoscale
  
  ### GCE
  - Decrease the number of GCE Read Requests made while deleting nodes
  - Base pricing of custom instances on their instance family type
  - Add pricing information for missing machine types
  - Add pricing information for different GPU types
  - Ignore the new `topology.gke.io/zone` label when comparing groups
  - Add missing stable labels to template nodes
  

### Internal

* 🔧 Update sponsors, remove Flint. PR [#10349](https://github.com/tiangolo/fastapi/pull/10349) by [@tiangolo](https://github.com/tiangolo).
* 🔧 Rename label "awaiting review" to "awaiting-review" to simplify search queries. PR [#10343](https://github.com/tiangolo/fastapi/pull/10343) by [@tiangolo](https://github.com/tiangolo).

  
    - Traffic to a service IP now starts getting rejected (as opposed to
      merely dropped) as soon as there are no longer any *usable*
      endpoints, rather than waiting until all of the terminating
      endpoints have terminated even when those terminating endpoints
      were not being used.
  
    - Chains for endpoints that won't be used are no longer output to

- Static pods were taking extra time to be restarted after being updated.  Static pods that are waiting to restart were not correctly counted in `kubelet_working_pods`. ([#116995](https://github.com/kubernetes/kubernetes/pull/116995), [@smarterclayton](https://github.com/smarterclayton)) [SIG Node]

- In the `wait.PollUntilContextTimeout` function, if `immediate` is true, the condition will now be invoked before waiting, guaranteeing that the condition is invoked at least once and then wait a interval before executing again. ([#119762](https://github.com/kubernetes/kubernetes/pull/119762), [@AxeZhan](https://github.com/AxeZhan))

- In `client-go` `SharedInformerFactory` will now support waiting for goroutines during shutdown.
   ([#112200](https://github.com/kubernetes/kubernetes/pull/112200), [@pohly](https://github.com/pohly))

of stuck finalizers. finalizers is a shared field, any actor with permission can reorder it. If the finalizer list is processed in order, then this can lead to a situation in which the component responsible for the first finalizer in the list is waiting for a signal (field value, external system, or other) produced by a component responsible for a finalizer later in the list, resulting in a deadlock. Without enforced ordering finalizers are free to order amongst themselves and are not vulnerable...

of stuck finalizers. finalizers is a shared field, any actor with permission can reorder it. If the finalizer list is processed in order, then this can lead to a situation in which the component responsible for the first finalizer in the list is waiting for a signal (field value, external system, or other) produced by a component responsible for a finalizer later in the list, resulting in a deadlock. Without enforced ordering finalizers are free to order amongst themselves and are not vulnerable...