Operationalizing AI Workloads on Kubernetes With OpenKruise - Zhang Zhen & Vec Sun
About this talk
This talk covers the operational challenges of handling AI workloads on Kubernetes, presented by Zhang Zhen from Alibaba Cloud and Vec Sun from Xiaohongshu. The speakers discuss the need for pre-warming container images to ensure quick startups due to large models and libraries, as well as the complexities of managing distributed training jobs that use PodGroups. They introduce OpenKruise's solutions, including cron-based pre-warming, an advanced disruption policy for maintaining PodGroup availability, and enhancements to support planned in-place restarts of entire PodGroups. These innovations aim to improve the reliability and efficiency of AI workload operations on Kubernetes at scale.
More from this event
See all 436 talks →
Best of KubeCon + CloudNativeCon Amsterdam 2026
2:17
The Quiet Work of Forever: Sustaining Open Source Communities - O. Hope Amaechi-Okorie, JSON Schema
26:24
Evolving KServe: The Unified Model Inference Platform for Both Predictive and... F. Spolti & J. Lee
32:40
Preventing S3 Cost Storms: Applying Cortex’s Efficiency Lessons to I/O-Heav... A. Fishman-Lichterman
5:32