Root Cause Analysis Method Based on Large Language Models with Residual Connection Structures

Fuente: arXiv
Salvato in:
Dettagli Bibliografici
Autori principali: Zhou, Liming, Liu, Ailing, Liu, Hongwei, He, Min, Zhang, Heng
Natura: Preprint
Pubblicazione: 2026
Soggetti:
Accesso online:
Tags: Aggiungi Tag
Nessun Tag, puoi essere il primo ad aggiungerne!!
_version_ 1866912891079229440
author Zhou, Liming
Liu, Ailing
Liu, Hongwei
He, Min
Zhang, Heng
author_facet Zhou, Liming
Liu, Ailing
Liu, Hongwei
He, Min
Zhang, Heng
contents Root cause localization remain challenging in complex and large-scale microservice architectures. The complex fault propagation among microservices and the high dimensionality of telemetry data, including metrics, logs, and traces, limit the effectiveness of existing root cause analysis (RCA) methods. In this paper, a residual-connection-based RCA method using large language model (LLM), named RC-LLM, is proposed. A residual-like hierarchical fusion structure is designed to integrate multi-source telemetry data, while the contextual reasoning capability of large language models is leveraged to model temporal and cross-microservice causal dependencies. Experimental results on CCF-AIOps microservice datasets demonstrate that RC-LLM achieves strong accuracy and efficiency in root cause analysis.
format Preprint
id arxiv_https___arxiv_org_abs_2602_08804
institution arXiv
publishDate 2026
record_format arxiv
spellingShingle Root Cause Analysis Method Based on Large Language Models with Residual Connection Structures
Zhou, Liming
Liu, Ailing
Liu, Hongwei
He, Min
Zhang, Heng
Artificial Intelligence
Root cause localization remain challenging in complex and large-scale microservice architectures. The complex fault propagation among microservices and the high dimensionality of telemetry data, including metrics, logs, and traces, limit the effectiveness of existing root cause analysis (RCA) methods. In this paper, a residual-connection-based RCA method using large language model (LLM), named RC-LLM, is proposed. A residual-like hierarchical fusion structure is designed to integrate multi-source telemetry data, while the contextual reasoning capability of large language models is leveraged to model temporal and cross-microservice causal dependencies. Experimental results on CCF-AIOps microservice datasets demonstrate that RC-LLM achieves strong accuracy and efficiency in root cause analysis.
title Root Cause Analysis Method Based on Large Language Models with Residual Connection Structures
topic Artificial Intelligence
url https://arxiv.org/abs/2602.08804