OkHttp is an HTTP client that’s efficient by default:

 * HTTP/2 support allows all requests to the same host to share a socket.
 * Connection pooling reduces request latency (if HTTP/2 isn’t available).
 * Transparent GZIP shrinks download sizes.
 * Response caching avoids the network completely for repeat requests.

OkHttp perseveres when the network is troublesome: it will silently recover from common connection

   certain eviction scenarios, leading to dropped response headers.

## Version 1.5.1

_2014-03-11_

 * Fix 1.5.0 regression where connections should not have been recycled.
 * Fix 1.5.0 regression where transparent Gzip was broken by attempting to
   recover from another I/O failure.
 * Fix problems where spdy/3.1 headers may not have been compressed properly.
 * Fix problems with spdy/3.1 and http/2 where the wrong window size was being

        },
        "content": "<div class=\"dashboard-header text-center\">\n<span>SERVICE: $service</span>\n</div>",
        "mode": "html"
      },
      "pluginVersion": "10.1.5",
      "transparent": true,
      "type": "text"
    },
    {
      "datasource": {
        "type": "prometheus",
        "uid": "${datasource}"
      },
      "fieldConfig": {
        "defaults": {
          "color": {

        },
        "content": "<div class=\"dashboard-header text-center\">\n<span>WORKLOAD: $workload.$namespace</span>\n</div>",
        "mode": "html"
      },
      "pluginVersion": "10.1.5",
      "transparent": true,
      "type": "text"
    },
    {
      "datasource": {
        "type": "prometheus",
        "uid": "${datasource}"
      },
      "fieldConfig": {
        "defaults": {
          "color": {

For traffic to unknown addresses, or to workloads that are not a part of the mesh, the traffic will just be passed through as is.
To make ztunnel more transparent, the original source IP address will be spoofed.
Additionally, `splice` will be used to make this proxying more efficient when possible.

For traffic in the mesh, things are a bit more complex:

        "mode": "markdown"
      },
      "pluginVersion": "10.1.5",
      "title": "Performance Dashboard README",
      "transparent": true,
      "type": "text"
    },
    {
      "collapsed": false,
      "datasource": {
        "type": "prometheus",
        "uid": "${datasource}"
      },
      "gridPos": {
        "h": 1,

    assertContent("a", getResponse(newRequest("/")))

    // Give the server time to disconnect.
    Thread.sleep(500)

    // This connection will need to be recovered. When it is, transparent gzip should still work!
    assertContent("b", getResponse(newRequest("/")))
    assertThat(server.takeRequest().sequenceNumber).isEqualTo(0)
    // Connection is not pooled.

### Network

- Ingress objects are now persisted in etcd using the networking.k8s.io/v1beta1 version ([#77139](https://github.com/kubernetes/kubernetes/pull/77139), [@cmluciano](https://github.com/cmluciano))
- Transparent kube-proxy restarts when using IPVS are now allowed. ([#75283](https://github.com/kubernetes/kubernetes/pull/75283), [@lbernail](https://github.com/lbernail))

- **Overhauled metrics**: Kubernetes has previously made extensive use of a global metrics registry to register metrics to be exposed. By implementing a metrics registry, metrics are registered in more transparent means. Previously, Kubernetes metrics have been excluded from any kind of stability requirements.

* Node-Problem-Detector configuration is now decoupled from the Kubernetes release on GKE/GCE. ([#73288](https://github.com/kubernetes/kubernetes/pull/73288), [@wangzhen127](https://github.com/wangzhen127))
* [IPVS] Allow for transparent kube-proxy restarts ([#75283](https://github.com/kubernetes/kubernetes/pull/75283), [@lbernail](https://github.com/lbernail))