Uncover Kubernetes cost-optimization trends
summary
CAST AI's 2024 benchmark analyzes 4,000 clusters across AWS Azure and GCP and finds typical CPU utilization painfully low (most clusters use under 20% of requested CPU, memory about 20%), driven by copy‑pasted requests, wrong instance families, and oversized dev clusters. Concrete wins include a 96% saved on Dev example ($25k → $1.5k) and automated spot arbitrage that produced a $1.5M windfall for one customer; recommendations: runtime automation, broad instance selection, pod disruption budgets, and Kubernetes in‑place resizing 1.30.
AI-written summary · about
chapters (13)
- 0:00Kubernetes cost optimization intro
- 1:07Industry resource waste
- 3:29Over-provisioning causes
- 6:45Pricing arbitrage with spot
- 9:34Automation for cost savings
- 13:27Shifting left vs runtime
- 15:55Scaling with cluster autoscaler
- 19:04Stateful vs stateless workloads
- 23:13Handling spot interruptions
- 25:45Kubernetes 1.30 right sizing
- 27:50Graviton adoption strategies
- 29:59Pod disruption budgets
- 33:08Q&A
related talks
33mKubernetes cost optimization - How to actually do it
26mDescubra las tendencias de optimización de costos en Kubernetes
14mCut Costs, Not Performance: Implement Kubernetes Cost Optimization Strategies
14mHow to cut your Kubernetes cloud costs by over 50%
33mThe Big Kubernetes Mistake Paying More Than Necessary
44mFinOps at scale: Lessons from scaling Kubernetes workloads
43mKubernetes cost optimisation - How to actually do it
40mHow an AI driven approach reduces cloud cost and makes your Kubernetes infrastructure autonomous