Keynote
summary
Argues that GitOps (using Argo and Flux) becomes the single source of truth so clusters and apps are reproducible and repairable; Weaveworks ships tools like eksctl, an EKS controller, Flux and Flagger plus Terraform and Kubernetes to manage full stacks. Demonstrates a real recovery that took 45 minutes and claims the system can scale to thousands of clusters. Shows automated progressive delivery controlling canaries across 15 clusters with metric-driven rollbacks, and describes model-based profiles for machine learning, web and mobile stacks.
AI-written summary · about
chapters (9)
- 0:00Welcome and thanks
- 0:42Introduction to GitOps
- 3:53Cluster recovery story
- 8:07GitOps evolution and benefits
- 10:59Reproducible clusters with GitOps
- 13:45Model-based deployment
- 16:20GitOps controller workflow
- 17:43GitOps cluster awareness
- 19:09Progressive delivery with GitOps
related talks
26mKeynote Session: GitOps as an Evolution of Kubernetes
06mKeynote: The Future of GitOps
06mKeynote: The Future of GitOps
35mKeynote: After Six Years, Is GitOps Boring?
26mGitOps: yea or nay?
31mLeveraging GitOps to wrangle your clusters and projects
36mKEYNOTE: Taking a step back and leveraging GitOps to wrangle your clusters and projects
06mKeynote: Beyond Automation: Kubernetes Success Requires a GitOps Mindset