In ROSA HCP, the hypershift.openshift.io/force-upgrade-to annotation is used for all OCP upgrades (z-stream/y-stream). Although this annotation allows control plane upgrades to start, even when the `ClusterVersionUpgradeable` status condition on the HostedCluster is `False`, it does not guarantee that the upgrade will succeeed.    

Tested a control plane upgrade from 4.15.5 --> to 4.15.6, but I believe the specific versions are not important.

100%    

What I did:
    1. Create a ROSA HCP cluster (not the latest version so that there are available upgrades)
    2. Initiate a control plane upgrade
    3. Remove the cluster's VPC's private subnet's route from `0.0.0.0/0 -> NAT Gateway` to simulate zero worker nodes
    4. Observe how the control plane upgrade progresses

What I think will work in general:
    1. Create a HyperShift cluster
    2. Remove all nodepools/worker nodes from the cluster
    3. Attempt to perform a control plane upgrade
    4. Observe how the control plane upgrade progresses     

The control plane upgrade is unable to progress, clusteroperators that depend on running pods on the worker nodes will be forever in a progressing state:

---
From `oc get clusterversion version -oyaml`
---
  - lastTransitionTime: "2024-04-17T19:40:10Z"
    message: |-
      Multiple errors are preventing progress:
      * Cluster operator image-registry is updating versions
      * Cluster operator ingress is degraded
      * deployment openshift-cluster-samples-operator/cluster-samples-operator is Progressing=False: ProgressDeadlineExceeded: ReplicaSet "cluster-samples-operator-75f4888746" has timed out progressing.
      * deployment openshift-console-operator/console-operator is Progressing=False: ProgressDeadlineExceeded: ReplicaSet "console-operator-77cfbc674b" has timed out progressing.
      * deployment openshift-insights/insights-operator is not available MinimumReplicasUnavailable (Deployment does not have minimum availability.) or progressing ProgressDeadlineExceeded (ReplicaSet "insights-operator-55846d47d8" has timed out progressing.)
      * deployment openshift-kube-storage-version-migrator-operator/kube-storage-version-migrator-operator is not available MinimumReplicasUnavailable (Deployment does not have minimum availability.) or progressing ProgressDeadlineExceeded (ReplicaSet "kube-storage-version-migrator-operator-775b779cd7" has timed out progressing.)
      * deployment openshift-monitoring/cluster-monitoring-operator is Progressing=False: ProgressDeadlineExceeded: ReplicaSet "cluster-monitoring-operator-b6cc85c5d" has timed out progressing.
      * deployment openshift-service-ca-operator/service-ca-operator is Progressing=False: ProgressDeadlineExceeded: ReplicaSet "service-ca-operator-c7b4dbb55" has timed out progressing.
    reason: MultipleErrors
    status: "True"
    type: Failing    

The control plane upgrade is able to "complete" - open to discussing the definition of "complete"