Microservices Data Mining for Analytics Feedback and Optimization

Abstract

When microservices-based architectures are adopted for an enterprise application, a basic requirement would be an evaluation of the performance with the objective of continuous monitoring and improved efficiency. This evaluation helps businesses obtain a quantitative measure of the benefits of a shift from monolith to microservices. Additionally, the metrics obtained could be used as a mechanism for continuous improvement of production application. This research proposes a model based on the principles of data mining called stream analytics feedback and optimization (SAFAO), which can be used to achieve a continuous optimization of microservices. Stream analytics is due to the fact that the analysis is performed on online application with continuously generated lived data. This approach has been tested in a simulated production environment based on Docker containers. The authors were able to establish empirical measures which were continuously extracted via a data mining methodology and then fed back into the running application through configuration management. The results show a continuous improvement in the performance of the microservices as indicated in the results presented in this research.