Sailing Ray Workloads with KubeRay and Kueue in Kubernetes
Talkat KubeCon + CloudNativeCon + Open Source Summit China 2023 · on Wednesday, 2023-09-27· at 11:00· 3层 307会议室| 3F Room 307·
summary
Explains running Ray workloads on Kubernetes using KubeRay (the Ray operator) to create and manage Ray clusters and Ray Job custom resources, and using Kueue (a Kubernetes-native job queuing system) to handle concurrent-job challenges such as resource allocation, multi-tenancy, fair sharing and preventing job starvation. Describes ByteDance’s production setup where thousands of jobs are submitted daily, debugging on long‑running clusters and launching regular jobs via Ray Job CRs, and how Kueue integrates to improve scheduling and resource management.
AI-written summary · about
related talks
26mIntroduction to Distributed ML Workloads with Ray on Kubernetes
24mSkyRay: Seamlessly Extending KubeRay to Multi-Cluster Multi-Cloud Operation
19mScaling Frontier AI with Ray on Kubernetes
28mRay on Public-Cloud k8s - experiments, lessons & best practices
28mScaling AI & Machine Learning Workloads With Ray on AWS, Kubernetes, & BERT
32mAn Overview of CloudKitchens's Ray-Powered ML Platform
32mRay in Production: Engineering Lessons from Spotify's Hendrix Framework
31mNext-Generation Data Science Workflows Using Ray