The Frugal k8s Engineer: Strategies for Amazon EKS
summary
Breaks down where EKS dollars go: control plane is fixed at $0.10/hour, data plane drives cost (CPU, memory, EBS, network, load balancers). Shows KubeCost using Prometheus metrics and AWS Cost & Usage data for namespace/pod attribution, and AWS Split Cost Allocation to split EC2/EBS across workloads. Demonstrates scaling a deployment from 7 to 125 replicas while Carpenter provisions nodes in seconds, using EC2 Fleet to balance spot vs on‑demand and recommending right‑sizing and consolidation for measurable savings.
AI-written summary · about
chapters (10)
- 0:00Introduction and Session Overview
- 0:49Understanding Amazon EKS
- 3:16Cost Components of Amazon EKS
- 5:35Cost Optimization Strategies
- 7:48Introduction to KubeCost
- 8:16KubeCost Demo
- 10:15AWS Split Cost Allocation Data
- 16:14Auto Scaling Strategies
- 18:55Carpenter in Action
- 25:52Key Takeaways and Conclusion
related talks
29mAmazon Elastic Container Service for Kubernetes (EKS): A case study - M. Re Ferre' and G. Carabelli
42mOptimize Amazon EKS workloads for performance and cost
32mNavigating the Sea of Kubernetes Costs: Kubecost as a Key Ally
33mKubernetes cost optimization - How to actually do it
35mKubernetes on AWS: Zero to hero
47mBuilding sustainable architectures with Kubernetes
17mKeynote: Kubernetes: Lessons from the Trenches
43mKubernetes cost optimisation - How to actually do it