DevOps World | Jenkins World 2018

Our project at GoDaddy uses continuous integration/continuous delivery to deploy each pull request, which means all 17,000,000 of our customers see each change. The CI/CD pipelines build across React, React Native and Node.js services. The initial pipeline - built without best practices in place - grew over time, until it became "the beast" that seemed impossible to tame. There were many different failure points and it wasn't clear which ones should be attacked first. This talk will explore the lessons we learned along the way (good and bad), early attempts at monitoring, some best practices and how the author built a Jenkins plugin to monitor our pipelines directly to gather the data needed to finally tame the beast.