The Real Python Podcast Michael Kennedy: Managing Your Own Python Infrastructure
85 snips
Oct 31, 2025 
Michael Kennedy, the creator of the Talk Python podcast and author of 'Talk Python in Production,' shares his expertise on deploying Python applications with autonomy. He covers how to avoid costly cloud lock-ins and discusses the benefits of using containerization and a single powerful VM for managing multiple applications. Michael also emphasizes the importance of opting for stack-native solutions over cloud-native complexities, and he reveals practical strategies for low latency and simple hosting setups. Tune in for insights that can elevate your Python deployment skills!
AI Snips
Chapters
Transcript
Episode notes
Prefer Portable Infrastructure
- Avoid vendor lock-in by preferring simple, portable infrastructure over cloud-specific managed services.
- Choose tools you can move between providers so you retain autonomy and reduce long-term costs.
Running 27 Apps On One VM
- Michael runs multiple real production apps (podcasts, courses, blog) on one cloud VM with Docker for isolation.
- He describes choosing Quart, MongoDB and Docker Compose to host 27 apps on a single eight‑CPU machine.
Use Docker Compose For Simplicity
- Use Docker and Docker Compose to isolate apps and keep deployments reproducible and inspectable.
- Run docker-compose logs and docker exec to debug and restart individual services without touching others.