Services or capabilities described in AWS documentation might vary by Region. To see the differences applicable to the AWS European Sovereign Cloud Region, see the AWS European Sovereign Cloud User Guide.

Quota and scheduling behavior for Ray workloads

Task Governance accounts quota at RayCluster granularity. A cluster reserves quota for its full declared size when it is admitted, and holds that quota for its whole lifetime regardless of load. A long-lived cluster that sits idle still counts against your team quota until you delete it.

Gang scheduling

Task Governance admits the head and all workers together. A RayCluster is not admitted until quota exists for the entire declared size, so a partially scheduled cluster never starts.

Preemption

When higher-priority work needs capacity, Task Governance preempts a lower-priority Ray workload. Preemption deletes the whole cluster, including the head.