Measuring Continuous Delivery by Steve Smith

Author:Steve Smith [Steve Smith]
Language: eng
Format: epub
Publisher: leanpub.com
Published: 2016-11-04T00:00:00+00:00

6. The Build Throughput Indicator

Introduction

A service is in a releasable state when it can be built repeatably and reliably. Every mainline commit should trigger a build server process that compiles the code, executes a comprehensive suite of automated tests, and publishes the deployable artifacts to an artifact repository as a release candidate.

A slow build process will create long feedback loops, which will discourage people from testing their changes pre-commit and monitoring the build server post-commit. Over time this can create a vicious cycle of declining build throughput and stability. On-demand production deployments will never be possible if the build process is not fast enough to regularly produce release candidates.

In A Practical Approach to Large-Scale Agile Development59, Gary Gruver et al describes how the HP LaserJet Firmware division accelerated build throughput. In 2008, 400 engineers worldwide worked on ten million lines of code, with a one week build time and a six week testing cycle. As a result HP was unable to satisfy business demand for several years, so invested in an automated build platform with extensive automated testing. By 2012, build time was three hours with a 24-hour testing cycle, and the business benefits were incredible. Time on new features increased by 700%, programmes in development increased by 140%, and development costs decreased by 40%.

A Build Throughput indicator can help a team position itself for Continuous Delivery, by uncovering build issues that prevent on-demand testing of release candidates.

Download

Copyright Disclaimer:
This site does not store any files on its server. We only index and link to content provided by other sites. Please contact the content providers to delete copyright contents if any and email us, we'll remove relevant links or contents immediately.