Kubernetes controller manager may fail to manage deployment rollout with minReadySeconds resulting into pod restarts.

We have a customer who is facing the issue as described in link[1]:
[1] https://github.com/kubernetes/kubernetes/issues/108266

The pending state is normal during the rollout. But the rollout should still proceed once the previous pod is terminated.

They leverage "minReadySeconds" to help improve availability of cluster control plane components during updates. However, they had to remove the use of "minReadySeconds" for many of their OpenShift components.

They need availability and speed and hence requesting for a fix. It's causing trouble to OpenShift control plane pods. They want the issue to be fixed.

Linked Referenced:
- https://kubernetes.io/docs/concepts/workloads/controllers/deployment/#min-ready-seconds
- https://github.com/openshift/ibm-roks-toolkit

OCP 4.10

Always

Steps are in the GitHub Link

Deployment not working as expected

Deployment should work as expected.