Context:

There's multiple scenarios where a cluster might not be upgradable:

• The CVO decides is not upgradable e.g https://github.com/openshift/enhancements/pull/828/files, or others where operators might signal non-upgradability.

• The API server decides is not upgradable based on deprecated fields.

DoD:

We should honour this decision and watch this condition when a new release image intent is expressed in the HC. Then we should communicate non-upgradability this in the HC conditions.

This could also be used to signal release changes that would break kube skew policy.

We should allow forcing an upgrade even if the status says it's not upgradable