* Aggregated used ports at the NodeInfo level for PodFitsHostPorts predicate. ([#42524](https://github.com/kubernetes/kubernetes/pull/42524), [@k82cn](https://github.com/k82cn))

- Fixed [121094](https://github.com/kubernetes/kubernetes/issues/121094) by re-introducing the readiness predicate for `externalTrafficPolicy: Local` services. ([#121116](https://github.com/kubernetes/kubernetes/pull/121116), [@alexanderConstantinescu](https://github.com/alexanderConstantinescu))

### Bug or Regression

- Fix 121094 by re-introducing the readiness predicate for externalTrafficPolicy: Local services. ([#121116](https://github.com/kubernetes/kubernetes/pull/121116), [@alexanderConstantinescu](https://github.com/alexanderConstantinescu)) [SIG Cloud Provider and Network]

- Removed the recently re-introduced schedulability predicate ([#109706](https://github.com/kubernetes/kubernetes/pull/109706)) as to not have unschedulable nodes removed from load balancers back-end pools. ([#111691](https://github.com/kubernetes/kubernetes/pull/111691), [@alexanderConstantinescu](https:/...

### Bug or Regression

- Fix 121094 by re-introducing the readiness predicate for externalTrafficPolicy: Local services. ([#121116](https://github.com/kubernetes/kubernetes/pull/121116), [@alexanderConstantinescu](https://github.com/alexanderConstantinescu)) [SIG Cloud Provider and Network]

- Integrated volume limits for in-tree and CSI volumes into one scheduler predicate. ([#77595](https://github.com/kubernetes/kubernetes/pull/77595), [@bertinatto](https://github.com/bertinatto))

## Deprecations and Removals

- API

  - The following APIs are no longer served by default:

### Bug or Regression

- Fix 121094 by re-introducing the readiness predicate for externalTrafficPolicy: Local services. ([#121116](https://github.com/kubernetes/kubernetes/pull/121116), [@alexanderConstantinescu](https://github.com/alexanderConstantinescu)) [SIG Cloud Provider and Network]

    yield
    # Clean up the ML models and release the resources
    ml_models.clear()


app = FastAPI(lifespan=lifespan)


@app.get("/predict")
async def predict(x: float):
    result = ml_models["answer_to_everything"](x)
    return {"result": result}
```

**Note**: This is the recommended way going forward, instead of using `startup` and `shutdown` events.

* Fixes and minor cleanups to pod (anti)affinity predicate ([#45098](https://github.com/kubernetes/kubernetes/pull/45098), [@wojtek-t](https://github.com/wojtek-t))

* kubeadm: Enable kubelet client certificate rotation ([#52196](https://github.com/kubernetes/kubernetes/pull/52196), [@luxas](https://github.com/luxas))
* Scheduler predicate developer should respect equivalence class cache ([#52146](https://github.com/kubernetes/kubernetes/pull/52146), [@resouer](https://github.com/resouer))