Creating a CI/CD Pipeline for Your Shared Libraries
summary
Capital One runs tens of thousands of pipelines and implements an automated CI/CD pipeline for a single Jenkins shared library that releases multiple times per day. The system uses the Jenkins shared library plugin plus a bootstrap custom-marker file, unit tests with Jenkins Pipeline Unit and self-run integration tests, DataDog metrics and Slack bots for alerts, a 30% canary rollout, and a fast rollback mechanism where deleting a release moves latest back one tag (delete GitHub release to rollback).
AI-written summary · about
chapters (8)
- 0:00Start
- 4:06extend your pipeline
- 4:20create a new git repository
- 16:23write a unit test
- 18:51perform integration tests
- 19:44executing an actual dinkus file with library functions against the library
- 28:02set up a data dog monitor
- 36:11set a custom marker file
related talks
26mFrom 100 Pipelines to 1
39mRemove your Testing Bottleneck with CI/CD Pipelines
12mContinuous Deployment in a heavily regulated environment
29mFrom scheduled builds to commit triggered pipelines
34mContinuous Delivery with Jenkins Workflow
14mHow we manage over1500 git repo’s CI/CD with a full InnerSource solution
34mTDD for jenkins pipelines
25mDevOps at Capital One: Focusing on Pipeline and Measurement