Making an Internal Kubernetes Offering Generally Available
Talkat KubeCon + CloudNativeCon North America 2019 · on Tuesday, 2019-11-19· at 16:25· Room 30ABCDE - San Diego Convention Center, Upper Level·
summary
Two years of work yield an internal GKE offering that moves 33% of eligible stateless backends via a self-service path: 15 production clusters with about 2,500 nodes, cluster lifecycle managed with terraform, deployments routed through existing tugboat and a sticky-assignment service compass. GA requirements include observed scalability and reliability, automated DIRT tests for cluster deletion and master failure, smoke-deploy metrics, cost and migration dashboards, namespace onboarding via a cookiecutter + SRM flow, canary parity, and staffed support shifts.
AI-written summary · about
chapters (18)
- 0:00Intro
- 0:48About Me
- 3:07Old Runtime
- 4:36GA Phase: Golden Path
- 6:09Multicluster Strategy: Management
- 9:26Multicluster Strategy: Compass
- 9:50Multicluster Strategy: Deployment
- 10:44Operational Metrics/Alerts
- 15:03Cost Tracking
- 15:53Business Impact
- 17:02Self-Service Experience: Namespaces
- 19:02Self-Service Experience: Apps
- 21:50Documentation
- 23:22Shared Support
- 26:16Feature Parity
- 27:15Our GA Requirements
- 27:56Future Plans
- 29:28Migration & GA Learnings
related talks
35mKubernetes journey
22mAnatomy of a production grade Kubernetes cluster
23mMore Power, Less Pain: Building an Internal Platform with CNCF Tools
37mRunning Kubernetes in a Manufacturing Line
29m11 5 years of kubernetes and looking to the future
36mMulti-Cluster Kubernetes Quick Delivery with IaC and GitOps
14mLessons from the Kubernetes Adventure
24mSelf-Service Multi-Cloud K8s Platform at a Regulated Financial Enterprise