# Knowledge-aware Alert Aggregation in Large-scale Cloud Systems: a Hybrid Approach

**Authors:** Jinxi Kuang, Jinyang Liu, Junjie Huang, Renyi Zhong, Jiazhen Gu, Lan Yu, Rui Tan, Zengyin Yang, Michael R. Lyu  
**arXiv:** [2403.06485](https://arxiv.org/abs/2403.06485)  
**Date:** 2024-03-11  
**Summary page:** https://www.summarizepaper.com/en/arxiv-id/2403.06485v1/

## Key points

- Authors propose COLA, a novel hybrid approach for alert aggregation in large-scale cloud systems
- COLA leverages external knowledge in the form of Standard Operation Procedure (SOP) to supplement existing methods
- COLA combines correlation mining and Large Language Model (LLM) reasoning for efficient online alert aggregation
- Experimental results show that COLA outperforms state-of-the-art methods while maintaining comparable efficiency levels
- Authors deployed COLA in their real-world cloud system named Cloud X
- Research accepted by Proceedings of the 46th International Conference on Software Engineering: Software Engineering in Practice (ICSE SEIP 2024)
- Available for further reading through DOI link 10.1145/3639477.3639745 or PDF link http://arxiv.org/pdf/2403.06485v1

## Summary

The authors propose COLA, a novel hybrid approach for alert aggregation in large-scale cloud systems. This approach leverages external knowledge in the form of Standard Operation Procedure (SOP) to supplement existing methods and combines correlation mining and Large Language Model (LLM) reasoning for efficient online alert aggregation. The experimental results demonstrate that COLA outperforms state-of-the-art methods while maintaining comparable efficiency levels. Additionally, the authors share their experience deploying COLA in their real-world cloud system named Cloud X. This research was accepted by Proceedings of the 46th International Conference on Software Engineering: Software Engineering in Practice (ICSE SEIP 2024) and is available for further reading through DOI link 10.1145/3639477.3639745 or PDF link http://arxiv.org/pdf/2403.06485v1.