Stability and code maintenance

This book mentions many plugins and a of languages and testing tools. It is OK to experiment in development and then push to acceptance, but the more diversity you have in production, the more skills are needed to maintain and especially to write a fluent workflow. Subtle choices, such as pinning Jenkins plugins at known versions and keeping the production version of your Jenkins server stable for fixed periods, help with up-time. Just as importantly, monitoring the load and offsetting most of the jobs away from the master Jenkins ensues a high degree of determinism in the timing of the jobs.

To limit job maintenance implies keeping configuration simple and similar. This is not realistic in a complex organization with a high degree of diversity. Using a test-driven approach helps; conventions also simplify configuration. As the diversity increases, communicating and agreeing to the conventions becomes important. Simple strategies such as one source of documentation...