Jenkins Scaling and Organization for an Efficient CI
summary
Camunda describes an evolution from bare-metal Hudson to VMs to Docker and Kubernetes, ending with a cloud-backed Jenkins fleet that runs about 2,500 jobs and produced ~22,000 builds last month. The stack uses Docker/Docker Swarm early on, moves to Kubernetes plugin on GCP, employs job-dsl and a homegrown Groovy generator plus the configuration as code plugin, and reports much faster builds (minutes instead of hours) and improved maintainability.
AI-written summary · about
chapters (13)
- 0:00Intro
- 2:21ClaCamunda - Scale with the company
- 4:11CI 1.0 - The Age of Virtual Machines
- 7:00CI 2.0 - The Age of Containers
- 13:42CI 2.0 - Problems
- 17:16Our philosophy for Cl3
- 20:00Jenkins - Kubernetes plugin
- 21:51Jenkins - Configuration as Code Plugin
- 25:21How does it work: Environments
- 27:19How does it work: Jobs
- 28:38What's the situation today?
- 30:12How does it work: Autonomy
- 31:57Conclusions
related talks
41mScaling Jenkins
35mOptimizing Your CI: Lessons Learned from a Successful Jenkins Rebuild
55mContinuous Integration on Steroids
13mEvolution of containers in continuous integration
23mMoving from Jenkins to Jenkins X: Scaling and Accelerating CI/CD
42mJenkins: Enabling Agility and Efficiency
06mBuilding Continuous Delivery Pipelines
05mJenkins in the Cloud