[1] Proposed title of this feature request
-> Request to change PodTopologySpread defaultConstraints.

[2] What is the nature and description of the request?
-> 

Customer do maintenance on one rack at a time so we need to ensure that Pods are spread across Racks.

https://kubernetes.io/docs/concepts/scheduling-eviction/topology-spread-constraints/#internal-default-constraints  says:
If you don't configure any cluster-level default constraints for pod topology spreading, then kube-scheduler acts as if you specified the following default topology constraints:

defaultConstraints:
  - maxSkew: 3
    topologyKey: "kubernetes.io/hostname"
    whenUnsatisfiable: ScheduleAnyway
  - maxSkew: 5
    topologyKey: "topology.kubernetes.io/zone"
    whenUnsatisfiable: ScheduleAnyway

This means that Pods are spread across both hostnames and zones, but the default maxSkew values are really high so Deployment/StatefulSets with 3 replicas (as is common) can still be scheduled on the same Node or rack.

To resolve this using vanilla Kubernetes we could create a similar config with maxSkew=0:

apiVersion: kubescheduler.config.k8s.io/v1
kind: KubeSchedulerConfiguration
...
profiles:
  - schedulerName: default-scheduler
    pluginConfig:
      - name: PodTopologySpread
        args:
          defaultConstraints: []
          defaultConstraints:
            - maxSkew: 0
              topologyKey: "kubernetes.io/hostname"
              whenUnsatisfiable: ScheduleAnyway
            - maxSkew: 0
              topologyKey: "topology.kubernetes.io/zone"
              whenUnsatisfiable: ScheduleAnyway
 

How would we do the same with OpenShift?

 [3]Why does the customer need this? (List the business requirements here)
-> 

It is important that all of the Pod replicas of a Deployment/StatefulSet run on different Nodes so that there is no service outage with Node maintenance.
We do maintenance on one rack at a time so we need to ensure that Pods are spread across Racks.

List any affected packages or components.
->