Automated Regression Testing Framework for Microservice Based Enterprise Applications

Abstract

Microservice-based enterprise applications require fast regression testing because changes in one service can affect APIs, databases, message flows, configurations, and connected services. This study presents an automated regression testing framework that combines change detection, dependency analysis, risk-based test selection, containerized execution, and feedback from earlier test runs. The framework identifies direct and indirect service impact and selects contract, integration, workflow, deployment, and smoke tests according to the detected change. It was evaluated against full-suite testing and dependency-only selection across increasing microservice change scope. The proposed method maintained fault detection above 95% while reducing normalized execution time compared with complete regression testing. It also detected more indirect failures than dependency-only selection because it considered historical execution evidence and service criticality. The framework can shorten pipeline feedback time, reduce unnecessary testing cost, and support safer continuous delivery in large enterprise microservice environments without sacrificing meaningful regression coverage during frequent production release cycles.