- The kube-controller-manager will set the `observedGeneration` field on pod conditions when the `PodObservedGenerationTracking` feature gate is set. ([#130650](https://github.com/kubernetes/kubernetes/pull/130650), [@natasha41575](https://github.com/natasha41575)) [SIG API Machinery, Apps, Node, Scheduling, Storage, Testing and Windows]

- Graduated `PodObservedGenerationTracking` feature to beta and had it on by default. This feature meant that the top level `status.observedGeneration` and `status.conditions[].observedGeneration` fields in Pods were populated to reflect the `metadata.generation` of the podspec at the time that the status or condition was reported. ([#132912](https://github.com/kubernetes/kubernetes/pull/132912), [@natasha41575...

* DaemonSet controller actively kills failed pods (to recreate them) ([#40330](https://github.com/kubernetes/kubernetes/pull/40330), [@janetkuo](https://github.com/janetkuo))
* DaemonSet ObservedGeneration ([#39157](https://github.com/kubernetes/kubernetes/pull/39157), [@lukaszo](https://github.com/lukaszo))

### Deployment

- `kubectl wait` ensures that observedGeneration >= generation to prevent stale state reporting. An example scenario can be found on CRD updates. ([#97408](https://github.com/kubernetes/kubernetes/pull/97408), [@KnicKnic](https://github.com/KnicKnic))

### Documentation

- Fix k8s.io/apimachinery/pkg/api/meta.SetStatusCondition to update ObservedGeneration ([#95961](https://github.com/kubernetes/kubernetes/pull/95961), [@KnicKnic](https://github.com/KnicKnic)) [SIG API Machinery]

* Rebase hyperkube image on debian-hyperkube-base, based on debian-base. ([#48365](https://github.com/kubernetes/kubernetes/pull/48365), [@ixdy](https://github.com/ixdy))
* change apps/v1beta2 StatefulSet observedGeneration (optional field) from a pointer to an int for consistency ([#49607](https://github.com/kubernetes/kubernetes/pull/49607), [@dixudx](https://github.com/dixudx))