Introduction

1. What This Module Covers

Here is the industry's uncomfortable statistic: most outages are self-inflicted. Not hardware, not traffic, not attackers. Changes. A deploy, a config push, a migration. The system was fine until someone improved it.

The reliability module taught you to survive the world's failures. This module teaches you to survive your own releases, and to see what is happening while you do.

Five patterns, two jobs. The first two are about observing. They tell you whether each server should receive traffic, and they follow one request through thirty services