Unlocking the Power of Machine Learning for Cloud Workload Opt.
summary
Spot uses ML on cloud provider telemetry and VM-level signals to predict spot instance interruptions and drive Kubernetes actions: scoring markets, shifting instance types/zones, and right-sizing nodes and pods. The system claims up to 80% savings by preferring low-risk markets and adapting autoscaling and vertical sizing, learns from thousands of customers to build a market score that precedes observed interruptions, and integrates into products like Spot Ocean for automated cluster decisions.
AI-written summary · about
chapters (5)
- 0:00Introduction and overview
- 1:24Cost reduction vs optimization
- 2:30Key cost optimization aspects
- 7:50Machine learning for optimization
- 10:10Q&A and closing
related talks
31mAccelerate AI adoption: automated cloud efficiency
47mReliable, secure, and efficient cloud operations
39mHow automating cloud operations accelerated time to market at NU
21mHow you can scale up with Amazon EC2 for up to 90% less
16mCracking the code: AI workloads at scale on AKS
10mOptimizing Machine Learning (ML Ops) in Financial Services with Kubernetes and Cloud Technologies.
21mLessons Learned Running GKE Clusters on Spot Instances
42mAWS Spot instances: huge savings (or huge risks?)