Deployment Process

Deployments to the example environment follow a fixed sequence: build, test, deploy to a lower environment, verify, then promote the same build artifact to production.

The same artifact that was tested in a lower environment is the one deployed to production — nothing is rebuilt between environments. This avoids a class of bugs where a rebuild introduces a difference between what was tested and what actually ships.

Production deployments require an approval step from a second engineer before they proceed, and the deployment pipeline automatically rolls back if health checks fail within a defined window after the new version starts serving traffic.