For hypershift clusters, individual node pools can be set to 0 replicas if the sum of nodes is at least 2 for the entire cluster.  Currently, this is caught via an error from the backend api call when we scale the node counts.

We need to add logic, on the frontend, that prevents the user from setting node counts less than 2 for the entire cluster.

Acceptance Criteria:

• Have pool size stored somewhere convenient like redux (possibly)
• On save of the Edit Node Count modal, the sum of replicas will be verified to be at least 2 across all pools.
• When only one node pool available, we should probably make the minimum choice 2
• Confirm Rosa wizard machine pool step won't let user go below 2 machine pools