Research on the efficiency of implementing a microservice system for monitoring users and resources on the AWS cloud platform

Abstract

The relevance of this study is driven by the need to improve the effectiveness of monitor-ing microservice-based systems in cloud environments for e-commerce, particularly on the AWS platform. This is achieved through a comparative analysis of implementations developed in different programming languages and an assessment of their suitability for use in high-load systems. Special attention is given to ensuring the stability and performance of system components, optimizing the utilization of computing and network resources, increasing the speed of user request processing, and improving the efficiency of resource management under dynamically changing load conditions. The paper considers key aspects of studying the effec-tiveness of the microservice system for monitoring users and resources of the AWS cloud plat-form, implemented using C# and Go. The results of performance and resource utilization be-tween implementations in the C# and Go programming languages were analyzed to determine the feasibility of using each language in e-commerce systems with a cloud architecture. To reproduce a load close to real user behavior, a PHP script was developed. It provides paral-lel execution of HTTP requests and collection of service response statistics. A series of load tests were performed with a gradual increase in the number of simultaneous requests. The results obtained allow to analyze the use of processor resources, RAM, and service response time at different load levels. Comparison of average values showed a stable difference be-tween implementations, which is confirmed by both tabular data and graphical monitoring results in the AWS EC2 environment. Additional analysis of long-term moderate load metrics showed that in the absence of significant differences in network traffic and the number of packets, the increased CPU utilization in the C# implementation is due to the peculiarities of code execution and the request processing model. Thus, the conducted study confirmed that the implementation of the metrics microservice on Go demonstrates lower computational re-source consumption and shorter response time under the same load conditions, which justifies the feasibility of its use when building microservice monitoring systems in cloud environ-ments.