Part 1 of 7 · 1 min read
Why this page comes fourth
In this module you will
- Describe the four stages of a delivery pipeline and what each one checks.
- Explain canary releases and blue-green deployments, and say when each fits.
- Plan a rollout across instances and regions, with bake time between steps.
- Split a backward-incompatible change into three safe steps.
- Explain feature flags as a way to separate deployment from release (common practice, not from the books).
Most outages begin with a change. Modules 1 to 3 gave the signals that show whether a change hurt users. This page uses those signals to release changes in small, reversible steps. Sources: "Understanding Distributed Systems" (UDS), chapter 19, and "Building Microservices" (BM), chapter 7. Numbers marked "made up" only show the idea.