Receive daily AI-curated summaries of engineering articles from top tech companies worldwide.
Endigest AI Core Summary
Kubernetes v1.35 introduces workload aware scheduling with new APIs and features to improve multi-Pod workload placement efficiency.
•The new Workload API (scheduling.k8s.io/v1alpha1) lets users define pod groups and apply scheduling policies, such as gang scheduling with a minCount threshold
•Gang scheduling enforces all-or-nothing placement: Pods wait at a Permit gate until the full group (minCount) can be scheduled simultaneously, preventing partial resource consumption
•If only a subset of the group is scheduled within a 5-minute timeout, all Pods are rejected and returned to the queue to free resources
•Opportunistic batching (Beta, enabled by default) speeds up scheduling of identical Pods by reusing feasibility calculations without requiring explicit opt-in
•
Future roadmap includes workload-level preemption, multi-node DRA topology-aware scheduling, autoscaling integration, and multi-workload scheduling simulations
This summary was automatically generated by AI based on the original article and may not be fully accurate.
Kubernetes
Google Cloud
AWS