Solving configuration drift using GitOps and ArgoCD
summary
A demo on GKE shows a three-node Kubernetes cluster where an app scaling change and a manual kubectl edit that breaks traffic are detected and corrected by ArgoCD. ArgoCD is run in pull‑based GitOps mode (default autosync polling ~every 3 minutes) with self‑healing enabled to immediately revert configuration drift; advice includes using tagged Git revisions for production and restricting direct kubectl access so the GitOps automation remains authoritative.
AI-written summary · about
chapters (15)
- 0:00Introduction
- 4:53Configuration drift explained
- 7:06Push vs pull deployment
- 10:42GitOps principles
- 13:12Argo CD custom resources
- 15:03Sync policy and self healing
- 17:34Argo CD UI demo
- 23:15Self healing in Argo CD
- 25:47Avoid manual cluster changes
- 26:50Automatic drift correction demo
- 28:18Argo CD limits and security
- 30:24Q&A
- 33:55GitOps repo structure
- 36:04Helm integration
- 37:27Summary
related talks
39mGitOps — How to Marry Git and Kubernetes for Deployment and Operations
30mManage your kubernetes cluster in the GITOPS way !!
29mDemystifying Gitops with ArgoCD
27mImplementing GitOps with ArgoCD: A Practical Guide
31mLeveraging GitOps to wrangle your clusters and projects
42mLet's Learn About ArgoCD
34mImplementing GitOps on Kubernetes: ArgoCD
25mGitOps and ArgoCD: Natura’s Evolutionary Journey to Orchestrate Multiple K8S Clusters