Natural Language Summarization Enables Multi-Repository Bug Localization by LLMs in Microservice Architectures

TL;DR

This paper introduces a natural language reasoning approach for bug localization in microservice architectures, transforming codebases into hierarchical summaries and performing NL-to-NL search, outperforming traditional retrieval methods.

Contribution

It presents a novel NL-based bug localization method that uses hierarchical summaries and a two-phase search, improving accuracy and interpretability over existing code retrieval techniques.

Findings

Achieves Pass@10 of 0.82 and MRR of 0.50 on DNext dataset

Outperforms retrieval baselines and agentic RAG systems like GitHub Copilot

Provides transparent NL-based search path for bug localization

Abstract

Bug localization in multi-repository microservice architectures is challenging due to the semantic gap between natural language bug reports and code, LLM context limitations, and the need to first identify the correct repository. We propose reframing this as a natural language reasoning task by transforming codebases into hierarchical NL summaries and performing NL-to-NL search instead of cross-modal retrieval. Our approach builds context-aware summaries at file, directory, and repository levels, then uses a two-phase search: first routing bug reports to relevant repositories, then performing top-down localization within those repositories. Evaluated on DNext, an industrial system with 46 repositories and 1.1M lines of code, our method achieves Pass@10 of 0.82 and MRR of 0.50, significantly outperforming retrieval baselines and agentic RAG systems like GitHub Copilot and Cursor. This work demonstrates that engineered natural language representations can be more effective than raw source code for scalable bug localization, providing an interpretable repository -> directory -> file search path, which is vital for building trust in enterprise AI tools by providing essential transparency.