Performance Analysis of a CodeIgniter 4-Based Web Stock Management System Using Locust

System performance needs to be evaluated to determine its ability to handle simultaneous user requests. An increase in the number of virtual users may affect the system's response time and stability. Therefore, load testing is required to measure system performance under varying user loads. This study aims to analyze the performance of a CodeIgniter 4-based web stock management system using Locust and to identify performance changes as the number of virtual users increases. This research employs a quantitative approach through several stages, including test environment preparation, test scenario design, test execution, data collection, and result analysis. Testing was conducted using 2, 4, 6, 8, and 10 virtual users, with a duration of 60 seconds for each scenario. The analyzed parameters include the number of requests, failed requests, average response time, median response time, P95, P99, Current Requests per Second (RPS), and Failure Rate. The test results show that the number of requests increased from 76 with 2 virtual users to 320 with 10 virtual users. The average response time also increased from 199.97 ms to 430.95 ms. All test scenarios recorded zero failed requests, resulting in a Failure Rate of 0%. The highest P95 value was recorded in the 10-virtual-user scenario at 920 ms. Based on the evaluation criteria established in this study, namely a maximum P95 of 1,500 ms and a Failure Rate of less than 5%, all test scenarios met the specified criteria. These results indicate that Locust recorded no failed requests during testing; however, an increase in the number of virtual users tended to be accompanied by an increase in system response time. Thus, Locust-based testing provides an overview of the system's performance in handling simultaneous user requests.