Health check endpoints! I just remembered this issue and had a thought: Perhaps if the deployment was split into dagit, daemon and user code(s) then the health check endpoint on dagit could also check the other two and return their status as well?