The challenge of migrating "operated" workloads with crane and RHO will quickly
become a reality that we have to contend with. We were able to kick this can
down the road with MTC because OLM was only tech preview with 3.11, but as
users begin to migrate from 4.x -> 4.x clusters, it is going to become
dramatically more common.

It's possible that this could be scoped to only solving the migration of OLM
operated workloads, which is potentially a more tractable problem, because it
would involve the migration of the OLM resources. Effectively a resubscription
on the target side, but there are a lot of problems that have to be considered
here.

Conceptually, it feels like the right way to migrate an operated workload is to
actually migrate the CR that causes an operator to instantiate the workload, and
to delegate the deployment/construction/config of the workload to the operator.

If we just try to wholesale lift-and-shift operated workload resources, you
quickly start to fight the operator as it continues to believe that it owns
the resources and we're unable to make changes to it.