[FEA][one-click] Validate N+1 compute cluster registration with one-click deployment

[sbaum1994]

Is your feature request related to a problem? Please describe.

The one-click quickstart supports a control plane plus GPU cluster layout, and the self-managed stack separates the control-plane lifecycle from the per-cluster NVCA operator lifecycle. We need explicit validation that adding an additional compute cluster after the first install works as expected.

This matters because a self-managed NVCF control plane should support more than one GPU cluster without clobbering registration state or reinstalling the control plane.

Describe the solution you'd like

Validate and document the N+1 cluster flow for one-click deployment.

The validation should cover:

• Install control plane and compute cluster A.
• Add compute cluster B using a separate kube context and cluster name.
• Confirm both clusters remain registered and healthy.
• Confirm functions can deploy to the expected cluster target.
• Confirm uninstalling cluster B does not remove cluster A or the control plane.

Describe alternatives you've considered

Operators can use the manual Helmfile and NVCA registration flow, but one-click is the preferred fresh-install path and should support the common multi-cluster expansion workflow.

Additional context

Suggested acceptance criteria:

• One-click flow succeeds for the first compute cluster.
• One-click or follow-up command succeeds for the second compute cluster.
• Cluster-specific registration values remain isolated.
• nvcf-cli self-hosted status shows both clusters correctly.
• Docs include the verified N+1 workflow and cleanup commands.