Mission Critical Kubernetes
summary
Built Kubernetes-on-Kubernetes using sub-clusters where each sub-cluster runs its own etcd petset and three HA master pods as a Deployment, enabling rolling upgrades by changing image tags. Infrastructure is provisioned with Terraform and configured with Ansible on CoreOS; CI tears down and recreates a full management+sub-cluster stack in ~30–35 minutes for end-to-end tests. Monitoring uses Prometheus+Sensu with namespace-label health checks; Vault integration and host-native networking replace Flannel. Design targets management cluster control of nodes and claims 100% functional uptime for banks, eCommerce and research customers.
AI-written summary · about
chapters (10)
- 0:00Intro
- 2:46Overview
- 3:22Existing Technology
- 4:57Cloud Configuration
- 6:21Ansible
- 13:20Demonstration
- 14:26CI Job
- 18:18Monitoring
- 23:07Next Steps
- 27:13Questions
related talks
36mBuilding a CI/CD driven infrastructure for managing Kubernetes clusters
37mManaging and Running Multiple Kubernetes Clusters in Hybrid Setups
30mEdge Kubernetes and Healthcare: A Next-Gen Architecture that Scales Across Thousands of Locations
36mManaging und Running Multiple Kubernetes Clusters
36mThe Architecture of a Multi-Cloud Environment with Kubernetes
53mManaging Containers at Scale with CoreOS and Kubernetes
35mKubernetes journey
44mEssential-8 Security in Practice with Kubernetes