Multicluster Batch Jobs Dispatching with Kueue at CERN

Learn about multicluster batch jobs dispatching through a keynote presentation from CERN's Lead Platforms Infrastructure and Google's Staff Software Engineer. Explore solutions for managing GPU demand across multiple clusters, regions, and clouds using Kueue. Discover how to automatically locate capacity, dispatch jobs, and monitor status in various environments including fixed-size on-premises setups, autoscaled cloud clusters, and hybrid configurations. Gain practical insights from CERN's implementation experience through a demonstration and valuable lessons learned during their deployment process.