Eliminate repetitive calls to rosa describe for a single cluster

[akrzos]

I believe we can make a fairly low effort reduction in calls to the ocm/rosa api simply by eliminating situations like the following:

2025-09-16 13:20:01,705 INFO Thread-2 hypershift - create_cluster: Cluster p4-alex1-0001 installation started on the 1 try
2025-09-16 13:20:01,705 INFO Thread-2 rosa - get_metadata: Getting information for cluster p4-alex1-0001
2025-09-16 13:20:01,705 DEBUG Thread-2 utils - subprocess_exec: rosa describe cluster -c p4-alex1-0001 -o json
2025-09-16 13:20:02,839 INFO Thread-2 hypershift - get_metadata: Getting information for cluster p4-alex1-0001
2025-09-16 13:20:02,839 DEBUG Thread-2 utils - subprocess_exec: rosa describe cluster -c p4-alex1-0001 -o json

We can see that the call to get_metadata actually generates two rosa describe commands.