Copper Dispatch packages an internal scheduler as a container. Its team is considering GKE because it needs Kubernetes APIs and control over workload scheduling. A cluster name on an architecture diagram does not settle how nodes connect to Google services, how Pods obtain credentials or which application endpoint is exposed. Those are separate design decisions with separate tests.
Start by storing the supplied private-node and workload-pool intent. Then review what it establishes and what remains untested. The browser creates a local cluster record. It does not start Kubernetes, schedule a Pod, connect a private node or issue a workload credential. Work through the same distinctions you would use when checking a real cluster rollout.
The private-node and identity settings are independent configuration choices.