How to Un-Flake Flaky Tests- a New Hire's Toolkit
summary
Inventory shows a suite spanning 14 microservices and Selenium Grid with ten nodes; practical fixes focus on concrete causes and tools: reproduce flaky runs, inspect videos/screenshots, use Chrome DevTools to test selectors and network payloads, set breakpoints and interspersed steps, tail Heroku logs, and run an HTML diffing script to capture DOM changes. Examples include a debounce bug where rapid star clicks register wrong value (6 vs 10) and a chromedriver version mismatch; aim for reliable waits, debug API payloads, and clear triage steps like tailing Heroku logs.
AI-written summary · about
chapters (12)
- 0:00Introduction and context
- 3:37Testing environment overview
- 5:47Defining flaky tests
- 8:08Selenium errors and debugging
- 15:54Selenium speed and debouncing
- 18:44Superuser perspectives
- 21:25Triage and stack traces
- 26:15Debugging techniques
- 34:39Diffing HTML changes
- 39:07System internals and logs
- 42:22Managing version compatibility
- 44:12CI environments and resources
related talks
48mFlaky Tests 2.0
01h 09mFlaky Tests: The Method
40mFinding Flakiness: The Real Reasons Your Tests Aren't Reliable
36mFixing Flaky Tests Like a Detective
22mTest Confidence: Types of Flake & How to Remediate
26mKeeping Your Tests Lean
46mFlaky, Fragile, and Forgotten: The Fall of a Test Automation Project
28mTest flakiness; methods for dealing with flaky tests