* Update to fluentd-gcp:1.28.3, rebased on ubuntu-slim:0.8 ([#43928](https://github.com/kubernetes/kubernetes/pull/43928), [@ixdy](https://github.com/ixdy))
* Fix polarity of NodePort logic to avoid leaked ports ([#43259](https://github.com/kubernetes/kubernetes/pull/43259), [@thockin](https://github.com/thockin))

- Kube-proxy: fix a bug on UDP `NodePort` Services where stale connection tracking entries may blackhole the traffic directed to the `NodePort` ([#98305](https://github.com/kubernetes/kubernetes/pull/98305), [@aojea](https://github.com/aojea))

  * Service.Spec.ExternalTrafficPolicy was 'service.beta.kubernetes.io/external-traffic' annotation.

  * Service.Spec.HealthCheckNodePort was 'service.beta.kubernetes.io/healthcheck-nodeport' annotation.

### **Node Components**
#### Container Runtime Interface

### Other notable changes

* Fix unintended change of Service.spec.ports[].nodePort during kubectl apply ([#24180](https://github.com/kubernetes/kubernetes/pull/24180), [@AdoHe](https://github.com/AdoHe))

- Kube-proxy: fix a bug on UDP NodePort Services where stale conntrack entries may blackhole the traffic directed to the NodePort. ([#98305](https://github.com/kubernetes/kubernetes/pull/98305), [@aojea](https://github.com/aojea)) [SIG Network]

### Bug or Regression

- Added new feature gate `ServiceNodePortStaticSubrange`, to enable the new strategy in the `NodePort` Service port allocators, so the node port range is subdivided and dynamic allocated `NodePort` port for Services are allocated preferentially from the upper range. ([#114418](https://github.com/kubernetes/kubernetes/pull/114418), [@xuzhenglun](https://github.com/xuzhenglun))

- The `kube-apiserver` will now expose four new metrics to inform about errors on the clusterIP and nodePort allocation logic. ([#120843](https://github.com/kubernetes/kubernetes/pull/120843), [@aojea](https://github.com/aojea))
- The `volume_zone` plugin will consider `beta` labels as `GA` labels during the scheduling

- Added support to the kube-proxy nodePortAddresses / --nodeport-addresses option to
  accept the value "primary", meaning to only listen for NodePort connections
  on the node's primary IPv4 and/or IPv6 address (according to the Node object).
  This is strongly recommended, if you were not previously using
  --nodeport-addresses, to avoid surprising behavior.

* Exclude nodes labeled as master from LoadBalancer / NodePort; restores documented behaviour. ([#44745](https://github.com/kubernetes/kubernetes/pull/44745), [@justinsb](https://github.com/justinsb))

## Before Upgrading