When deriving multiple physical models (e.g., PostgreSQL + MongoDB + OpenAPI) from a polyglot core, what governance strategies do you recommend to prevent long-term model drift between targets?
In particular:
Do you enforce all structural changes at the polyglot level only?
How do you handle target-specific optimizations without breaking conceptual alignment?
Has anyone implemented automated drift detection using the CLI?
I’m interested in enterprise-scale workflows where models evolve continuously across teams.