LEMMA-RCA: A Large Multi-modal Multi-domain Dataset for Root Cause Analysis

arXiv:2406.05375v4 Announce Type: replace Abstract: Root cause analysis (RCA) is crucial for enhancing the reliability and performance of complex systems. However, progress in this field has been hindered by the lack of large-scale, open-source datasets tailored for RCA. To bridge this gap, we introduce LEMMA-RCA, a large dataset designed for diverse RCA tasks across multiple domains and modalities. LEMMA-RCA features various real-world fault scenarios from Information Technology (IT) and Operational Technology (OT) systems, including microservices, water distribution, and water treatment systems, with hundreds of system entities involved. We evaluate the performance of six baseline methods on LEMMA-RCA across various settings, including offline and online models, as well as single and multi-modal settings. Our study demonstrates the utility of LEMMA-RCA in facilitating fair evaluation and promoting the development of more robust RCA techniques. The dataset is publicly available at https://www.lemmarca.info, with archival metadata and access instructions deposited under DOI: https://doi.org/10.5281/zenodo.20516735.

Related

• Root cause analysis via difference graph discovery from linear time-series data
• The Abstention Protocol: RCA for Clos Fabrics
• LLM-Based Automated Diagnosis Of Integration Test Failures At Google

Source: arXiv cs.AI | 2026-08-26