*    from the peer.
 *
 * Web sockets may fail due to HTTP upgrade problems, connectivity problems, or if either peer
 * chooses to short-circuit the graceful shutdown process:
 *
 *  * **Canceled:** the web socket connection failed. Messages that were successfully enqueued by
 *    either peer may not have been transmitted to the other.
 *

    fun ackSettings()

    /**
     * Read a connection-level ping from the peer. `ack` indicates this is a reply. The data
     * in `payload1` and `payload2` opaque binary, and there are no rules on the content.
     */
    fun ping(
      ack: Boolean,
      payload1: Int,
      payload2: Int,
    )

    /**
     * The peer tells us to stop creating streams. It is safe to replay streams with

James Wald <******@****.***> 1773565302 -0400

     * followed by those lines.
     *
     * HTTPS responses also contain SSL session information. This begins with a blank line, and then
     * a line containing the cipher suite. Next is the length of the peer certificate chain. These
     * certificates are base64-encoded and appear each on their own line. The next line contains the
     * length of the local certificate chain. These certificates are also base64-encoded and appear

- Adds alpha metrics `apiserver_peer_proxy_errors_total` and `apiserver_peer_discovery_sync_errors_total` to apiserver to track errors encountered in peer proxying and peer discovery ([#137065](https://github.com/kubernetes/kubernetes/pull/137065), [@richabanker](https://github.com/richabanker)) [SIG API Machinery]

              (System.nanoTime() - startNanos) / 1e9f, e);
        }
      }
    ```

### Timeouts ([.kt][ConfigureTimeoutsKotlin], [.java][ConfigureTimeoutsJava])

Use timeouts to fail a call when its peer is unreachable. Network partitions can be due to client connectivity problems, server availability problems, or anything between. OkHttp supports connect, write, read, and full call timeouts.

=== ":material-language-kotlin: Kotlin"

- Kubeadm: ignore EINVAL when unmounting /var/lib/kubelet peer mounts during reset ([#137569](https://github.com/kubernetes/kubernetes/pull/137569), [@fuweid](https://github.com/fuweid)) [SIG Cluster Lifecycle]

### Other (Cleanup or Flake)

Nikita Putikhin, Niklas Vangerow, Nitin Srinivasan, Oleg Shyshkov, Olli Lupton, Om Thakkar, Pankaj Kanwar, Parker Schuh, Paul Ganssle, Pauline Sho, Pavithra Eswaramoorthy, Pedro Gonnet, pemeliya, Penporn Koanantakool, Perry Gibson, Peter Buchlovsky, Peter Gavin, Peter Hawkins, Pham Binh, Phani Paladugula, Philipp Hack, Praneeth Mandala, Praveen Batra, psinfinity, Qingwei Zhang, Quentin Khan, Quoc Truong, QZero, Rachel Han, Raffi Khatchadourian, Ram Rachum, RasheedAli-Shaik, Raviteja Gorijala, Reed...

- Kube-scheduler: in Kubernetes 1.33, the number of devices that can be allocated per ResourceClaim was accidentally reduced to 16. Now the supported number of devices per ResourceClaim is 32 again. ([#131679](https://github.com/kubernetes/kubernetes/pull/131679), [@mortent](https://github.com/mortent)) [SIG Node]

- Kube-scheduler: in Kubernetes 1.33, the number of devices that can be allocated per ResourceClaim was accidentally reduced to 16. Now the supported number of devices per ResourceClaim is 32 again. ([#131662](https://github.com/kubernetes/kubernetes/pull/131662), [@mortent](https://github.com/mortent)) [SIG Node]