Application deployment: getting code from your laptop to production, and back out if you need to

Most deployment guides stop at the command that copies your code to a server. That's the part that rarely goes wrong. The trouble lives around it: a config value that exists on your laptop and nowhere else, a database migration that can't run twice, a release that works for you and breaks for users stuck on a cached old version.

This section covers the whole move from local to staging to production for web apps, APIs and content sites. We look at build artifacts, environment variables, migrations, zero-downtime cutovers, DNS changes and the checks you run in the first ten minutes after a release. Every guide ends the same way: how to confirm the deploy worked for a real user, and how to reverse it if it didn't.

Whether you ship to shared hosting, a VPS, a platform like Render or Fly.io, or a major cloud provider, the principles carry over. The tools change. The questions don't: what changed, how do we know it works, and how fast can we undo it?