fix: authorization controller triggers deletion of clusteradmin

[reshnm]

What this PR does / why we need it:

When the authorization component is being deleted it must be ensured that all ClusterAdmin resources with the same name as the authorization component are being deleted before the deletion can continue.
Otherwise the ClusterAdmin resource will be orphaned and can't be deleted.

Which issue(s) this PR fixes:
Fixes #112

Special notes for your reviewer:

Release note:

Ensure that the corresponding ClusterAdmin resource is deleted before the ManagedControlPlane is being deleted.

[Diaphteiros]

This fixes #112, I guess?

[Diaphteiros]

This fixes #112, I guess?

Hm, no, I guess it doesn't, because it removes only a single ClusterAdmin resource and not potentially more user-created ones.

Should we directly implement a fix for #112 ?
I assume that all ClusterAdmin resources belonging to an MCP become orphaned and unable to be deleted when the MCP is removed?

[reshnm]

This fixes #112, I guess?

Hm, no, I guess it doesn't, because it removes only a single ClusterAdmin resource and not potentially more user-created ones.

Should we directly implement a fix for #112 ? I assume that all ClusterAdmin resources belonging to an MCP become orphaned and unable to be deleted when the MCP is removed?

A MCP should only delete its own ClusterAdmin resources.
The Project Workspace Operator has a feature that can block deletion until all resources of a GVK are being deleted.

The owner of the Workspace should be able to delete all ClusterAdmins before the workspace is being deleted.