Guard: Scalable Straggler Detection and Node Health Management for Large-Scale Training
Abstract
Training frontier-scale foundation models involves coordinating tens of thousands of GPUs over multi-month runs, where even minor performance degradations can accumulate into substantial efficiency losses. Existing health-check mechanisms, such as NCCL tests or GPU burn-in, primarily focus on functional correctness and often fail to detect fail-slow behaviors that silently degrade system performance. In this paper, we present Guard, a scalable system for detecting stragglers and ensuring node health in large-scale training clusters. Guard combines lightweight online performance monitoring during training with an offline node-sweep mechanism that systematically evaluates and qualifies nodes before they participate in production workloads. This design enables Guard to detect both acute failures and long-running fail-slow behaviors that traditional diagnostics cannot capture. Deployed on large-scale foundation model pretraining workloads, Guard improves mean FLOPs utilization by up to 1.7x, reduces run-to-run training step variance from 20% to 1%, increases mean time to failure (MTTF), and significantly reduces operational and debugging overhead. These results demonstrate that proactive straggler detection and systematic node qualification are critical for maintaining stable and efficient large-scale training.
Cite
@article{arxiv.2605.17879,
title = {Guard: Scalable Straggler Detection and Node Health Management for Large-Scale Training},
author = {Guanliang Liu and Abhinandan Patni and Congzhu Lin and Zoe Zeng and Jack Wittmayer and Josh Wu and Ashvin Nihalani and Binxuan Huang and Yinghong Liu and Rory Na and Anthony Ko and Alexander Zhipa and Cong Cheng and Mi Sun and Vijay Rajakumar and Rejith George Joseph and Parthasarathy Govindarajen},
journal= {arXiv preprint arXiv:2605.17879},
year = {2026}
}
Comments
Proceedings of the 9 th MLSys Conference, Bellevue, WA, USA, 2026