I'm curious how other data engineers handle keeping up to data with schema migrations in downstream databases?