We currently have a dependency on the OCP version to uninstall HCP clusters. The control plane is bound to an OCP version and tears down the cluster and its cloud components. 

The downside is that when a cluster of version 4.15.1 is stuck uninstalling due to a bug in 4.15.1, we are not able to resolve the bug by fixing the uninstalling components (e.g. control plane operator), as the control plane won't be able to upgrade after and uninstallation is triggered.

This goal of this RFE is to enable us to push bug fixes that will unblock stuck installations.

Some ideas:

• decouple the uninstallation flow from the hosted control plane version (probably difficult architecture wise at this point)
• add an additional garbage collection process in the hypershift operator, this could clean up everything on the management cluster cloud resource side via labels and skip the cloud resource customer side deletion
• allow the hypershift operator to consume a configmap where known issues (such as log segments) can be selected to detect uninstallation bugs, and the remediation can be added as script