Large-scale kubernetes deployment, which tools to use and how - Rachid Zarouali
summary
Experience report on running hundreds of Kubernetes clusters using Cluster API versus Crossplane, showing a working pipeline that uses Argo CD to drive helmified cluster manifests and a seed-of-seeds pattern for scale. Demonstrates OpenStack provider quirks: slow image builds, undocumented options and a crippling Keystone rate limit, recommends flatcar images and intermediate seed clusters. Demo deploys six clusters from one seed, upgrades via manifest changes work automatically, hybrid worker/control-plane setups require forking providers.
AI-written summary · about
chapters (8)
- 0:00Introduction and context
- 2:08Project challenges
- 6:27Technology selection
- 10:34Understanding Cluster API
- 15:00Lessons and limitations
- 19:02Takeaways and future steps
- 23:03Demo and workflow
- 36:55Q&A session
related talks
32mDon't Start With 500 Clusters: Patterns in Scaling Multi-Cluster Kubernetes…
17mOur Journey to Kubernetes at Scale
34mInfrastructure at Scale Using ArgoCD
01h 38mScaling Multi-cluster Management with Argo CD and Application Sets
32mLaying the Foundation: Real World Kubernetes Deployment Patterns
38mInfrastructure deployments with ArgoCD HA and Crossplane
33mEliminating the Gap Between Development and Production with Kuber...
35mKubernetes journey