Laying the Foundation: Real World Kubernetes Deployment Patterns
summary
Compares managed, distribution and roll-your-own Kubernetes with concrete tradeoffs: managed services block control‑plane access so pod security policies may be impossible, distributions constrain unsupported features, and DIY requires upgrade and lifecycle plans. Recommends thinking multi‑cluster early, sizing around a 500‑node practical limit before deep tuning, choosing CNI (Calico vs others) and storage before deployment, using OIDC for auth, enforcing resource quotas, and using Velero backup tool and operators for safe extensions while avoiding premature service mesh adoption.
AI-written summary · about
chapters (7)
- 0:00Intro
- 2:22Managed Kubernetes Distribution Roll Your Own
- 8:03Yesterday, today, and tomorrow Back to the basics Enable your stakeholders
- 13:26Bedrock There are some things you won't change Container Networking Persistent Storage Connectivity
- 18:15Multi-team? Multi-tenant? Consistent AuthZ/Authi Policies, policies, policies Backup and restore
- 24:38Multi-cluster mindset That'F' word: Federation Resources
- 28:22Highly extensible Plan on extending Operators
related talks
31mHow NOT to Start with Kubernetes
36mKubernetes, Decisions
48mLeveraging the power of Kubernetes with patterns
41mArchitecting Kubernetes-Based Internal Developer Platforms: Essential Patterns...
33mKubernetes-Driven Multi-Cloud Deployments: Maximizing Scalability and Flexibility
31mKubernetes is Your Platform: Design Patterns For Extens...
01h 02mKubernetes Patterns
41mKubernetes the Very Hard Way