Desire: We should be opinionated to assess what we can migrate, and what we can’t migrate. Discover the things that are questionable, call them out, and handle the things that we know we can migrate.

• Can we prep the DVM infra prior to a migration so we can verify it's even possible and effectively front-load problems that may occur so it can be solved as part of initial evaluation of the product?

• Can we build the infrastructure to be able to run N pre-flight heuristics, and make it very easy to add as we learn? It's very challenging to figure out "what are the things that are questionable" when we don't know what we don't know.