How Jenkins Builds the Netflix Streaming Service
summary
Jenkins runs 1,600 active jobs and about 2,000 builds per day and orchestrates bake deploy via the bakery and Asgard to produce Netflix AMIs, publish artifacts to artifactory, and deploy into autoscaling groups. The setup uses a single master with AWS slaves, custom plugins and housekeeping system groovy scripts, a Job DSL for templated job definitions, and the Dino slave plugin for self-registration. Scaling pain points prompt a sharded master plan using EBS-backed job volumes and blue‑green style redeploys.
AI-written summary · about
chapters (17)
- 0:00Intro
- 0:18Overview
- 0:56Engineering Tools
- 2:34History on Jenkins
- 7:35Build and Deployment Pipeline
- 13:07Jenkins Use Cases
- 15:11Jenkins Usage
- 16:24Netflix Architecture
- 21:13Balancing Needs
- 22:32Sharded Architecture
- 23:44Deployment
- 26:45Jenkins Job
- 27:49Job dsl plugin
- 29:44Job dsl demo
- 31:00Job maintenance
- 32:19Groovy scripts
- 34:06Wrap up
related talks
35mJenkins Scaling and Organization for an Efficient CI
38mOperating Jenkins at Scale
45mNetflix’s Secret to Safe Automation at Scale
33mJenkins 101- 38mOperating Jenkins at Scale
44mScaling Your Jenkins Jobs
22mAll Things Jenkins and Cloud Foundry
34mHelping Enterprises Scale Their Jenkins Software Practices and Tools