How Google Cloud Hosts and Manages Kubernetes at Scale
summary
Google Container Engine claims to run Kubernetes at true production scale — spinning up many thousands of clusters in ~3–5 minutes via a single API/button by using prebuilt Container-Optimized OS images, managed MIG-backed node pools, and saved master credentials. It adds auto-upgrade and auto-repair beta for masters and nodes, pod and node autoscaling, Stackdriver logging/monitoring, and a federated ingress with a single global static IP. Pricing example: ~$0.15/hr per cluster (requires >6 nodes) and a 99.5% SLA.
AI-written summary · about
chapters (16)
- 0:00Introduction
- 4:07Disneyland
- 9:24Behind the Scenes
- 15:01Monitoring
- 18:40Node Pools
- 21:00Enabling Auto Upgrade
- 21:50Enabling Auto Repair
- 22:02Running Applications Without Container Engine
- 23:17Auto Scaling
- 24:42Federated Ingress
- 26:10How to use Federated Ingress
- 27:56Logging and Monitoring
- 28:47Rolling Out Kubernetes
- 30:44Reporting and Monitoring
- 31:15Cost
- 32:03Summary
related talks
44mContainers and CI/CD on Google Cloud
16mRunning Services at Google Scale
58mKeynote: So, I have all these Docker containers, now what?
35mKubernetes the easy way
24mHow Does Google Release Kubernetes in GKE
46mCloud Platform for DevOps Professionals
30mCI/CD for Hybrid and Multi-Cloud Customers
44mA Practical Guide to Container Scheduling