Microservice architecture has become one of the most widely adopted approaches in the design of modern software systems due to its support for the independent development, deployment, and scalability of system components. This architecture plays a significant role in enhancing system flexibility and facilitating DevOps processes. However, the absence of standardized design guidelines, the inherent complexity of inter-service communication, and poor initial design decisions often lead to the emergence of architectural anti-patterns. Such deficiencies can severely impact the performance, maintainability, and evolvability of microservice-based systems. Therefore, accurately identifying existing anti-patterns is a critical step toward improving architectural quality and preventing the long-term decline in system quality and maintainability of microservice systems. Although several studies have addressed the identification of anti-patterns in microservice-based systems, automated approaches for detecting architectural anti-patterns are still in their infancy and require further advancements to achieve practical effectiveness. In this context, we propose a fully automated approach capable of detecting anti-patterns within microservice systems. Using the proposed method, we identified 15 anti-patterns across 25 microservice-based systems. The results demonstrate that our approach can effectively detect microservice anti-patterns with an average precision of 93% and a recall of 97.9%. Hence, the proposed method can serve as a valuable tool to assist developers in improving their development practices.

Keywords