Reduce Costs and Optimize on Your Cloud with Karpenter
summary
Karpenter is shown provisioning Kubernetes nodes using NodeClass blueprints and the EC2 Fit API to pick among 800+ instance types (Spot, on‑demand, Savings Plans) for cost optimization. The talk demonstrates a real deployment cost drop (about $111k/month on‑demand vs ~$52k/month using Spot with diversification and consolidation) and describes Karpenter emitting node claim lifecycle events, obeying a consolidation policy to shrink idle capacity, and integrating with Graviton (ARM) and EventBridge interruption signals for safer Spot usage.
AI-written summary · about
chapters (16)
- 0:00Introduction
- 1:08Cloud cost reduction laws
- 3:34Karpenter vs Cluster Autoscaler
- 5:27Why Karpenter cuts costs
- 7:36AWS purchase options impact
- 8:18Graviton cost advantages
- 10:47Node definition strategies
- 12:32Karpenter labels and selectors
- 14:14Consolidation policies
- 16:27Handling Spot interruptions
- 20:15Sustainability and carbon footprint
- 22:04Demo: Karpenter on EKS
- 25:21Demo: Manual scaling
- 28:04Demo: Real-time cost monitoring
- 29:30Demo: On-Demand vs Spot
- 33:13Summary and savings
related talks
22mUse Karpenter to scale, optimize & upgrade K8s
22mKarpenter Efficient scaling of Kubernetes clusters
18mKarpenter : Your trusted Companion for Kubernetes Scaling
47mBuilding sustainable architectures with Kubernetes
29mCutting Costs and Boosting Efficiency with Karpenter and KEDA on Kubernetes
01h 01mJust-in-time Nodes for Any AWS EKS Cluster: Karpenter
27mHarness the power of Karpenter to scale, optimize & upgrade K8s
33mKarpenter: Best Practices and Cost Optimization Techniques