1. Proposed title of this feature request
Possibility to apply KubeSchedulerConfiguration to configure cluster-level constraints

2. What is the nature and description of the request?
In specific use-cases such as AWS Outpost or similar deployments, its desired to apply more/better control over application scheduling to achieve best possible availability and performance for application. While it can be achieved, application by application using Pod Topology Spread Constraints, it's still rather cumbersome as each application need to apply a default set of configuration to make it work. For OpenShift Container Platform 4 administrators, knowing the infrastructure well and application landscape well, it's rather desired to enforce such constrains by default using Cluster-level default constraints. While this could be achieved using the Secondary scheduler it's still not the desired way of achieving it as it will again require opt-in from the application.

Yet the goal is to make this work completely transparent for application and hence enforce it by default, using the default scheduler as per Cluster-level default constraints.

3. Why does the customer need this? (List the business requirements here)
In certain regulated and strict requirements, the OpenShift Container Platform 4 administrator teams knows the infrastructure and requirements best and hence would like to enforce certain constrains during scheduling of pods to achieve the overall availability and resilience of the application without the need for the application to think about it (unless, they need something different than the default). It's therefore requested to have a way to configure Cluster-level default constraints respectively to apply customization with regards to scheduling using Plugins that apply to multiple extension points.

4. List any affected packages or components.

• kube-scheduler