The Site Reliability Workbook

Chapter 15. Configuration Specifics

By Dave Cunningham and Misha Brukman with Christophe Kalt and Betsy Beyer

Managing production systems is one of the many ways SREs provide value to an organization. The task of configuring and running applications in production requires insight into how those systems are put together and how they work. When things go wrong, the on-call engineer needs to know exactly where the configurations are and how to change them. This responsibility can become a burden if a team or organization hasn’t invested in addressing configuration-related toil.

This book covers the subject of toil at length (see Chapter 6). If your SRE team is burdened with a lot of configuration-related toil, we hope that implementing some of the ideas presented in this chapter will help you reclaim some of the time you spend making configuration changes.

Configuration-Induced Toil

At the start of a project’s lifecycle, configuration is usually relatively lightweight and straightforward. You might have a few files in a data-only format like INI, JSON, YAML, or XML. Managing these files requires little toil. As the number of applications, servers, and variations increases over time, configuration can become very complex and verbose. For example, you might have originally “changed a setting” by editing one configuration file, but now you have to update configuration files in multiple locations. Reading such configuration is also hard, as important differences are hidden in a sea ...