Analisis Kinerja Sistem E-Learning Universitas Pamulang Menggunakan Load Testing Berbasis Apache JMeter

Abstract

E-learning systems have become the backbone of modern education, but suboptimal performance can hinder the teaching and learning process. This study analyzes the performance of Pamulang University's e-learning system to identify potential barriers and ensure service stability. The main problem is the lack of in-depth evaluation regarding the system's capability to handle high user loads. The research objective is to measure and evaluate the e-learning system's performance under varying loads and provide improvement recommendations. The load testing method using Apache JMeter was applied to simulate real usage scenarios by 100 to 1000 concurrent users, with testing parameters including response time, throughput, and transaction failure rate. Results show significant performance degradation starting at 300 concurrent users with an average response time of 1486.66 ms. At 500 concurrent users, response time surged to 8407.28 ms and throughput decreased drastically to 38.52 req/s. At 1000 concurrent users load, the system demonstrated inability to scale with response time of 10860.18 ms, 90% percentile of 78840.30 ms, and error rate of 3.13%. In conclusion, the critical performance point of the system lies around 300-500 concurrent users, which serves as a basis for Pamulang University to enhance infrastructure capacity and codebase to ensure the sustainability of reliable digital education services.