How GitOps Should I Be? Lessons from a Kinda-GitOps Environment
summary
Intuit runs a pragmatic hybrid GitOps setup: 45 Argo CD instances serving ~20,000 apps across ~200 clusters, with region failover that bypasses GitHub because the GitHub SLA 20 minutes plus 20 minutes Argo CD recovery is too slow, so a kubectl CronJob scales east/west for faster recovery. Jenkins pipelines keep existing imperative steps and end with an Argo CD CLI sync, bootstrapping uses local sync to seed GitLab, and rendered manifests are produced by Jenkins and committed to a prod branch for auditable deploys.
AI-written summary · about
chapters (6)
- 0:00Intro
- 1:01Definition of GitOps
- 2:43Region Evacuation
- 5:26Pipeline-Triggered Syncs
- 7:07Bootstrapping Clusters
- 10:19Rendered Manifests
related talks
43mGitOps at Scale with and Argo CD on
01h 19mTutorial: Everything You Need To Become a GitOps Ninja
31mGitOps, the Final Frontier: Proposing, Promoting, and Reverting
27mImplementing GitOps with ArgoCD: A Practical Guide
30mThe Objectives And The Principles Behind GitOps
05mLessons from a Large-Scale GitOps Implementation
31mLeveraging GitOps to wrangle your clusters and projects
06mGitOps - At a Glance