DEVOURED
How Netflix Simplified Batch Compute with Kueue
Netflix migrated its batch compute jobs to Kueue, enabling advanced resource sharing and preemption-based scheduling on its Titus container platform.
What: Netflix replaced its proprietary Compute Managed Batch (CMB) system with Kueue, a Kubernetes-native job queueing project. This migration allows for efficient tenant-based resource borrowing and ensures high-priority jobs can preempt lower-priority tasks.
Why it matters: Custom batch schedulers are costly to maintain and often lack the flexibility of community-supported Kubernetes native tools. Leveraging Kueue aligns Netflix with open-source standards while solving complex resource contention problems.
Decoder
- Kueue: An open-source Kubernetes-native job queueing system that manages resource quotas and scheduling for batch workloads.
- Preemption: The ability of a scheduler to stop or pause a lower-priority task to free up resources for a higher-priority task.
Original article
Netflix migrated millions of batch jobs from its homegrown Compute Managed Batch (CMB) system to Kueue, an open-source Kubernetes-native job queueing system, resulting in significantly improved resource utilization across its container platform Titus. The migration, which allowed operators to switch tenants with the click of a button, brought features like preemption-based fair sharing that let teams borrow idle reserved capacity from other tenants while ensuring high-priority workloads can still preempt lower-priority ones when needed.
Explore all 56 articles in today's digest
Fresh dev news, AI-digested daily. Never miss a trend at devoured.fyi