Research map / Multi-agent systems and games / LLM agents
RCAFlow: A Workflow-Informed Hierarchical Planning Multi-Agent System for Root Cause Analysis
Yufei Gao, Zhengong Cai, Bowei Yang
Read the paperCheck your idea against 137,217 papers
AI-extracted summary. Written by AI from the paper's full text. Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.
Proposes RCAFlow, a workflow-informed hierarchical planning multi-agent LLM system for root cause analysis of microservice failures.
Problem
Root cause analysis in complex microservice systems, where existing deep learning methods lack flexibility and interpretability and prior LLM-based approaches are limited.
Approach
A multi-agent system that combines workflow information with hierarchical planning to structure LLM-driven root cause analysis, as described by the title and the partial abstract; further details are not available from the abstract alone.
Measured effects reported in the paper
| What was changed | Where | Measured | Effect |
|---|---|---|---|
| add workflow knowledge base (ablation) | RCAFlow (DeepSeek-V3) · OpenRCA Bank | Correct accuracy (%): 12.50 -> 40.44 | Improved |
| add workflow knowledge base (ablation) | RCAFlow (DeepSeek-V3) · OpenRCA Telecom | Correct accuracy (%): 15.69 -> 47.06 | Improved |
| add Git-inspired branch coordinator (ablation) | RCAFlow (DeepSeek-V3) · OpenRCA Bank | Correct accuracy (%): 25.00 -> 40.44 | Improved |
| add Planner with semantic retrieval (ablation) | RCAFlow (DeepSeek-V3) · OpenRCA Bank | Correct accuracy (%): 29.41 -> 40.44 | Improved |
| add state-aware analyzer (ablation) | RCAFlow (DeepSeek-V3) · OpenRCA Bank | Correct accuracy (%): 34.56 -> 40.44 | Improved |
| add reflection module (ablation) | RCAFlow (DeepSeek-V3) · OpenRCA Bank | Correct accuracy (%): 33.82 -> 40.44 | Improved |
| add few extra SOPs to the workflow | RCAFlow (DeepSeek-V3) · OpenRCA Bank | Correct accuracy (%): 40.44 -> 43.38 | Improved |
| raise confidence threshold theta (ablation) | RCAFlow (DeepSeek-V3) · OpenRCA Bank | Correct accuracy (%): rise-then-fall; APL increases | Mixed |
| use RCAFlow instead of Flow-of-Action | GPT-4o · OpenRCA Bank | Correct accuracy (%): 27.94 -> 41.91 | Improved |
| use RCAFlow instead of mABC | RCAFlow (DeepSeek-V3) · OpenRCA Market | Correct accuracy (%): 20.95 -> 34.46 | Improved |
Closest papers by meaning
- DyFlow: Dynamic Workflow Framework for Agentic Reasoning (NeurIPS 2025 · similarity 0.86)
- Benchmarking Agentic Workflow Generation (ICLR 2025 · similarity 0.83)
- Flow: Modularized Agentic Workflow Automation (ICLR 2025 · similarity 0.83)
- GraphFlow: A Graph-Based Workflow Management for Efficient LLM-Agent Serving (ICML 2026 · similarity 0.83)
- WorkflowLLM: Enhancing Workflow Orchestration Capability of Large Language Models (ICLR 2025 · similarity 0.82)
- Optimize Once, Execute Fast: Latency-Aware Multi-Agent Workflow Learning for Recurrent Queries (NeurIPS 2026 · similarity 0.82)
Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.
