SELF-HEALING WEB AUTOMATION: AN EMPIRICAL COMPARISON OF TRADITIONAL SELENIUM FRAMEWORKS AND HEALENIUM STACK

Abstract

Web user interfaces and mobile applications change rapidly and even minor modifications in the element address or Document Object Model (DOM) can invalidate test locators, causing brittle, flaky test suites and unstable CI/CD pipelines. Self-healing automation frameworks aim to reduce this fragility by automatically recovering from locator breakage at runtime. This paper presents a practical approach for understanding self-healing web automation and its various methodology like – Healenium and Selenide. We describe how they work, provide code-level details and propose an experimental setup that compares traditional Selenium tests with healing tests against a self-healing stack on many dimensions: locator-related flakiness, maintenance effort, and test execution time. Experimentation results indicate that self-healing can reduce locator-related failures by 40–60% and maintenance effort by 30–50%, at the cost of minor execution overhead. We discuss threats to validity, limitations, and how this pattern generalizes with mobile and API automation.