ROSA has a pattern in which they create 480 MachineSets, each normally scaled to zero, but, occasionally scaled up to 1.

We don't typically have customers with so many unique MachineSets, and so it would be good to understand the use case behind this, and see what the differences are between the MachineSets.

Investigating how the migration controllers will work in this scale would be worthwhile, and hopefully identify scaling issues in our design.

We should aim to look at this before we GA

cewong@redhat.com and joelspeed talked through this at the CAPI F2F in July '24 so will likely have useful context