LLMDebug: Prompt-Engineered Large Language Models for Automated Root Cause Analysis in Microservices Architectures

Debugging failures in modern microservices architectures presents significant challenges due to distributed execution, complex inter-service dependencies, and massive log volumes. Traditional rule-based approaches struggle with novel failure patterns and require extensive manual expertise. This paper introduces LLMDebug, a novel framework leveraging prompt-engineered Large Language Models (LLMs) for automated root cause analysis in distributed systems. We propose a multi-stage prompting strategy that combines chain-of-thought reasoning, retrieval-augmented generation (RAG), and multi-agent collaboration to diagnose failures across microservices. Our framework achieves $\mathbf{8 9 . 3 \%}$ accuracy ($\mathbf{9 5 \%}$ CI: $\mathbf{8 7 . 9 - 9 0 . 7 \%}$) in identifying root causes across $\mathbf{1 2}$ common failure scenarios, reducing mean-time-to-resolution (MTTR) by $68 \%$ compared to traditional monitoring tools. Extensive evaluation on a realistic microservices testbed with 8 services demonstrates that LLMDebug outperforms state-of-the-art baselines including DeepLog, LogRobust, and MicroRCA by $\mathbf{1 2 . 9 - 3 1}$ percentage points. We provide comprehensive ablation studies showing the contribution of each prompting strategy and release our complete implementation, datasets, and evaluation framework for reproducibility.