Update on the question:

To provide more background, I was interested in setting up a test cluster with differences in configuration from our production one but I didn't want actual clients's VMs to be deployed to it or to restart on it in case of an HA event and having to rely on the API all the time isn't an option since the helpdesk uses the CloudPlatform interface to deploy VMs and it wasn't reasonable for me to expect they would learn to use the API.  Below is the solution.

When you add host in CloudPlatform, you can specific a "host tag".   Likewise, the service offering will allow you to do the same. CP MGMT server won't deploy a VM to a host using tag unless the service offering has the matching tag.  So:

1. Create new cluster

2. As you add hosts to the new cluster, add a host tag, i.e.: DEV

3. Create a service offering with the same host tag.

4. Deploy a new VM using the purposely-created service offering.

The above steps will allow you or your helpdesk staff to deploy VMs on different clusters, as needed.  Just make the Service offering descriptions clear.  There's many reasons for wanting to do this:

- Maybe some clusters are less overprovisioned

- Maybe oyu want a test cluster

- I'm sure I'm missing some...