Flaky Test Automation and Mitigating Test Crashes in Agile Releases

Abstract

Flaky tests—tests that fail non-deterministically without any changes in code—pose a serious threat to the reliability and efficiency of automated testing pipelines. They lead to wasted engineering hours, reduced confidence in CI/CD systems, and increased costs. This paper provides an in-depth analysis of the causes of flaky test automation and presents actionable strategies to detect, diagnose, and mitigate test flakiness and crashes. We also explore emerging tools, AI-assisted techniques, and empirical studies that highlight industry practices for maintaining robust and deterministic test suites.In risk management and software testing industry, prevention aims to stop an event from happening in the first place, while mitigation focuses on reducing the severity or impact of an event that has already occurred or is unavoidable. Prevention is proactive, trying to stop a problem before it starts, while mitigation is reactive, managing the consequences of a problem. In this research, we are broadly describing techniques of mitigation of occurred and detected flaky test failures and automation crashes.