Recap and Practice

Key takeaways

• Infrastructure is a ladder of five rungs — static hosting, serverless, PaaS (Platform as a Service — a host that runs your app so you don't manage the server), managed containers, and raw VMs — and you should pick the simplest rung that does the job.
• Prefer managed services: outsource the operations (servers, patches, scaling, backups) while keeping the understanding.
• Let AI write your deploy config and CI (Continuous Integration — an automated robot that runs your checks on every code push) workflow, but read it before you trust it.
• "Managed" is not "magic" — your data, your secrets, your bill, and your understanding stay your responsibility.
• Cap spend, confirm HTTPS, and know how to roll back before you launch.

Try it

Take a small project you already have (even a single HTML page) and deploy it to a static host. Then ask AI to generate a GitHub Actions workflow that redeploys on every push, read the generated YAML line by line, and confirm you can find the "roll back to previous deploy" button in your host's dashboard.

I have a static site in this repo. Walk me through deploying it to a
free static host, then generate a GitHub Actions workflow that
rebuilds and redeploys on every push to main. Explain each step of
the YAML, and tell me where in the dashboard I roll back a bad deploy.