mABC

Notice

We will release more data in this repository as soon as possible. Here is init version code and wait for our data to be updated after passing policy review.

News

Our paper is accepted by EMNLP 2024 (Findings)! 🎉🎉

Abstract

The escalating complexity of micro-services architecture in cloud-native technologies poses significant challenges for maintaining system stability and efficiency. To conduct root cause analysis (RCA) and resolution of alert events, we propose a pioneering framework, multi-Agent Blockchain-inspired Collaboration for root cause analysis in micro-services architecture (mABC), to revolutionize the AI for IT operations (AIOps) domain, where multiple agents based on the powerful large language models (LLMs) perform blockchain-inspired voting to reach a final agreement following a standardized process for processing tasks and queries provided by Agent Workflow. Specifically, seven specialized agents derived from Agent Workflow each provide valuable insights towards root cause analysis based on their expertise and the intrinsic software knowledge of LLMs collaborating within a decentralized chain. To avoid potential instability issues in LLMs and fully leverage the transparent and egalitarian advantages inherent in a decentralized structure, mABC adopts a decision-making process inspired by blockchain governance principles while considering the contribution index and expertise index of each agent. Experimental results on the public benchmark AIOps challenge dataset and our created train-ticket dataset demonstrate superior performance in accurately identifying root causes and formulating effective solutions, compared to previous strong baselines. The ablation study further highlights the significance of each component within mABC, with Agent Workflow, multi-agent, and blockchain-inspired voting being crucial for achieving optimal performance. mABC offers a comprehensive automated root cause analysis and resolution in micro-services architecture and achieves a significant improvement in the AIOps domain compared to existing baselines.

Overview

Overview of mABC. Overall pipeline encapsulates the flow from alert inception to root cause analysis within mABC.

Two distinct workflows of agent. ReAct answer involves an iterative cycle of thought, action, and observation until a satisfactory answer is reached, while responses are directly formulated based on the prompt provided following the direct answer.

Vote process on Agent Chain

Enviornment and Running

Driven by OpenAI Version

1. python enviornment:

pip install -r requirements.txt

2. define your OPENAI_API_KEY and task.

export OPENAI_API_KEY="sk-xxx"

3. run script follow your task, example:

python main/main.py

Driven by Other Version

Try to replace /Users/knediny/Desktop/mABC/utils/llm.py.

Notice

Feel free to cite us if you like mABC, and you can contact me by zwpride@buaa.edu.cn.

@article{zhang2024mabc,
  title={mABC: multi-Agent Blockchain-Inspired Collaboration for root cause analysis in micro-services architecture},
  author={Zhang, Wei and Guo, Hongcheng and Yang, Jian and Zhang, Yi and Yan, Chaoran and Tian, Zhoujin and Ji, Hangyuan and Li, Zhoujun and Li, Tongliang and Zheng, Tieqiao and others},
  journal={arXiv preprint arXiv:2404.12135},
  year={2024}
}