Productivity of Technology Optimum Configurations for Automated Testing of Rich Component Systems

Abstract

Introduction. The transition of the software industry to distributed microservice and cloud-native architectures has significantly increased the flexibility of information systems, but has led to the phenomenon of “environment drift”. Exhaustive testing of such ecosystems causes a “combinatorial explosion”, which, under continuous integration and continuous deployment (CI/CD) pipelines, leads to connection pool exhaustion and critical budget overruns on cloud computing (FinOps), making complete enumeration impossible.

The purpose of the paper. Conducting a performance review, experimental benchmarking, and verifying the computational stability of the developed prototype of information technology (in the format of CLI Middleware) for selecting optimal test configurations of multi-component systems under conditions close to industrial ones.

Methods. An information technology has been developed in the form of CLI Mid dleware based on Node.js using combinatorial methods (Pairwise Testing, Ortho gonal Arrays) and evolutionary computations (genetic algorithms). To completely prevent the blocking of the main execution cycle (Event Loop), strictly deterministic evolutionary search parameters have been implemented (30 generations, population of 50) and the early stopping mechanism (Early Stopping) has been disabled.

Results. Experimental benchmarking on synthetic test suites (108 and 243 configurations) confirmed the system`s ability to automatically compress the configuration space by 70.086.0%. An isolated mathematical validation module proved the preservation of 100% coverage of pairwise (2-wise) interactions. The recorded execution time of the computational core is 612 ms, making it ideally suited for use as a synchronous step in CI/CD pipelines.

Conclusions. The effectiveness of the developed information technology has been proven, successfully solving the “combinatorial explosion” problem and eliminating the risks of blocking the main cycle. Achieving an integral efficiency metric of 0.9725 confirms the absolute readiness of the CLI Middleware architecture for industrial integration into corporate pipelines, allowing significant optimization of infrastructure costs without compromising the quality of interaction defect detection.