Semi-automated generation of web app performance tests from end-to-end GUI-level tests with E2E-Loader

Performance testing is essential to ensure that web applications remain responsive and reliable under varying workloads. However, defining realistic workloads is a complex and time-consuming task, often requiring a deployed system and manual handling of data dependencies between requests. E2E-Loader is a tool designed to automate workload generation for performance testing by leveraging existing End-to-End (E2E) GUI-level test cases. By analyzing network traffic captured during GUI test execution, E2E-Loader extracts sequences of HTTP and WebSocket requests and identifies data dependencies, enabling the creation of realistic workloads without requiring access to production logs or manual correlation scripting. The tool includes a graphical interface for workload configuration and generates executable Apache JMeter scripts. A preliminary evaluation conducted in an industrial setting demonstrated that E2E-Loader can produce workloads comparable to those manually crafted by practitioners, with significantly reduced effort. Future work includes expanding empirical validation across multiple systems and teams, improving dependency detection through semantic-aware techniques, enhancing usability, and extending support for additional performance testing frameworks.