- Add supervisord.conf to run all workers + health app in single container
- Update Dockerfile to use supervisor as default CMD (K8s still overrides)
- Add .gitignore for Python artifacts
- Update .env.example and README with HF deployment info
- Fix migration script for asyncpg (split SQL statements)
- Configure connection pools from .env variables
- Add MongoDB connection pooling
This enables running the full booking allocation engine in HF Spaces
while maintaining separate pod deployments for K8s/production.
