Aligning DevOps and Microservice Architecture: Empirical Mapping, Taxonomy, and RAG-Based Decision Support

DevOps and microservice architecture (MSA) are central to modern software delivery, yet their conceptual alignment remains fragmented across the literature. This study empirically analyzes how DevOps practices correspond to MSA principles and how these relationships inform architectural and operational decisions. From 84 peer-reviewed sources (2016–2025), we extract and normalize over 200 DevOps–MSA concept pairs and organize them into a five-category taxonomy: deployment automation and orchestration, observability and runtime diagnostics, scalability and modularity, security and governance, and configuration and environment management. To operationalize these mappings, we develop a retrieval-augmented generation (RAG) framework that combines sentence-transformer embeddings with GPT-based reasoning to generate bidirectional concept recommendations. Quantitative evaluation shows that RAG outperforms lexical and embedding-only baselines by roughly 20 % in precision and ranking, while qualitative assessment confirms semantic accuracy and contextual relevance. Deployment automation and observability emerge as the most empirically supported categories, whereas governance-related links remain largely conceptual. Industrial case studies in software, banking, and insurance sectors demonstrate measurable improvements in release speed and developer satisfaction when the framework supports transformation planning. The complete dataset and open-source RAG prototype are released to facilitate replication and extension. Overall, the findings indicate that empirically grounded RAG enhances traceability and evidence-based decision-making between architecture and operations in complex software systems.