ternalEtcd` that can be used to configure the HTTP endpoints for etcd communication in v1beta4. This field is used to separate the HTTP traffic (such as /metrics and /health endpoints) from the gRPC traffic handled by Endpoints. This separation allows for better access control, as HTTP endpoints can be exposed without exposing the primary gRPC interface. Corresponds to etcd's `--listen-client-http-urls` configuration. If not provided, Endpoints will be used for both gRPC and HTTP traffic. ([#134...

- One new field "InternalTrafficPolicy" in Service is added.
  It specifies if the cluster internal traffic should be routed to all endpoints or node-local endpoints only.
  "Cluster" routes internal traffic to a Service to all endpoints.
  "Local" routes traffic to node-local endpoints only, and traffic is dropped if no node-local endpoints are ready.

  - Deletion of remote endpoints will be triggered only for terminated endpoints (those present in the old endpoints map but not in the new endpoints map), whereas previously it was also done for terminating endpoints. ([#124092](https://github.com/kubernetes/kubernetes/pull/124092), [@princepereira](https://github.com/princepereira)) [SIG Network and Windows]

      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
      iptables, saving a bit of memory/time/cpu. (#106373, @aojea) [SIG Network]

- Fixed a bug where improper storage and comparison of endpoints led to excessive API traffic from the endpoints controller ([#94935](https://github.com/kubernetes/kubernetes/pull/94935), [@damemi](https://github.com/damemi)) [SIG Apps, Network and Testing]

- Fix kube-apiserver not to create default/kubernetes service endpoints before it reports readiness via the /healthz and therefore is ready to serve requests. Also early during startup old endpoints are remove which might be left over from a previously crashed kube-apiserver. ([#74668](https://github.com/kubernetes/kubernetes/pull/74668), [@sttts](https://github.com/sttts))

- Fixed a bug where improper storage and comparison of endpoints led to excessive API traffic from the endpoints controller ([#94934](https://github.com/kubernetes/kubernetes/pull/94934), [@damemi](https://github.com/damemi)) [SIG Apps, Network and Testing]

      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

  using Endpoints.
  
  If you are using the aggregated API server feature, and you are writing out
  the endpoints for it by hand (rather than letting kube-controller-manager
  generate Endpoints and EndpointSlices for it automatically based on the
  Service definition), then you should write out an EndpointSlice object rather

- Endpoints resources created by the Endpoints controller now have a label indicating this.
  (Users creating Endpoints by hand _can_ also add this label themselves, but they ought