Reworking Your Team for Kubernetes
summary
Yahoo migrates large legacy services onto Kubernetes and reports concrete results and lessons: infrastructure tuning yields up to 6:1 host reduction for identical performance, a 1% canary rollout finds problems early, and the horizontal pod autoscaler handles real traffic spikes with minimal fuss. Operational tools include Athens authentication, a custom ATS routing plugin to avoid overlay networks, image scanning with Claire, Helm and Chef for deploys, and fixes like SSDs to resolve package-manager lock contention when many pods start.
AI-written summary · about
chapters (12)
- 0:00Introduction and context
- 2:24Migration strategy
- 3:20Kubernetes architecture
- 5:07Infrastructure challenges
- 6:52Technical hurdles
- 9:44Utilization benefits
- 11:51Operational best practices
- 14:05Building the team
- 16:53Deployment and maintenance
- 19:48Security and reliability
- 24:34Customizing the system
- 27:26Future goals
related talks
35mThe Oregon Trail to Kubernetes [I]
19mPains and gains of Kubernetes for Developers, story from experience
46mIn Search of the Perfect Cloud Native Developer Experience
25mWhy is Anyone Using Kubernetes Anyway?
28m11 tricks to improve productivity and reduce your frustration with Kubernetes
27mMaximizing Developer Productivity in Kubernetes
36mThe Road to More Usable Kubernetes
05mProject Lightning Talk: Meshery: Kubernetes without Yaml, Is It Possible?