Beyond Fault Localization: A Trajectory-Level Study of LLM Agents for Microservice Root Cause Analysis

📅 2026-08-21
📈 Citations: 0
Influential: 0
📄 PDF
🤖 AI Summary
研究通过轨迹级分析方法评估LLM代理在微服务根因分析中的表现,提出DiagGuard框架以提高诊断准确性。
📝 Abstract
Existing evaluations of automated root cause analysis (RCA) for microservices assess diagnostic performance mainly by endpoint correctness: whether a method localizes the responsible service. This criterion enables comparison but does not reveal the evidentiary basis of a diagnosis or the fault-propagation route connecting the source to observed symptoms, both of which an on-call site reliability engineer needs to judge whether action is warranted. We therefore treat RCA as an observable diagnostic process. Our trajectory-level framework evaluates agent executions against manually curated service-level fault-propagation paths. Applied to a public microservice RCA benchmark, it analyzes 3,500 diagnostic trajectories, characterizing where agents investigate and how they use retrieved telemetry. We find a disconnect between answer correctness and diagnostic quality: an agent may localize the fault source yet fail to reconstruct its propagation. Successful investigations stay on the fault-impact surface, act on retrieved evidence, and broaden their query repertoire as the search deepens. Failures arise when decisive evidence is omitted, retrieved evidence is misinterpreted, or unsupported inference substitutes for missing evidence. We operationalize this taxonomy as DiagGuard, a two-stage defense-in-depth architecture in which grounding surveys available observations before localization and verification audits the diagnosis against them. In an independent setting with a different model, benchmark, and service topology, DiagGuard raises Acc@1 from 43.5% to 52.5%. These results show that trajectory-level evaluation exposes limitations hidden by final-answer metrics and provides actionable guidance for improving automated RCA.
Problem

Research questions and friction points this paper is trying to address.

root cause analysis
microservices
diagnostic process
fault-propagation
evidentiary basis
Innovation

Methods, ideas, or system contributions that make the work stand out.

trajectory-level evaluation
evidentiary basis
fault-propagation path
DiagGuard
🔎 Similar Papers
No similar papers found.
💼 Related Jobs
No related jobs found.
Q
Qisheng Lu
The Chinese University of Hong Kong, Shenzhen, China
Aoyang Fang
Aoyang Fang
The Chinese University of Hong Kong, Shenzhen
Software testingAIopsRoot cause analysis
Junjielong Xu
Junjielong Xu
The Chinese University of Hong Kong, Shenzhen
AI4SE
J
Jin'ao Shang
Xi’an Jiaotong University, Xi’an, China
S
Songhan Zhang
The Chinese University of Hong Kong, Shenzhen, China
Yifan Yang
Yifan Yang
South China University of Technology
Neural renderingImage synthesisophthalmologic
X
Xiaochuan Yan
The Chinese University of Hong Kong, Shenzhen, China
Pinjia He
Pinjia He
Assistant Professor, The Chinese University of Hong Kong, Shenzhen
Software EngineeringAI4SESE4AIAIOps