Say 3 replica groups, 10 nodes per group, all gossiping.

I would like to write a test that focuses on high node counts within a single region, with large replica groups. This is a more realistic test as regions don’t talk to each other, so we can test how requests are distributed in a high-density scenario.