The Power of Atomic Automated Tests: Our experience of Improved Test Efficiency...
summary
Claims concrete wins: 7,000 Selenium tests in 40 minutes and 2,500 Android tests in 2 hours, plus a reported 40× speedup from 40 parallel runners. Techniques actually used: API-driven setup and cleanup, DB seeding instead of full snapshots, reading OTPs from the DB, setting auth tokens into localStorage, and small isolated tests to enable parallelism. Examples include code-level JS DOM manipulation and before/after cleanup with exact timings.
AI-written summary · about
chapters (14)
- 0:00Atomic tests intro
- 1:11Test automation benefits
- 3:17Atomic vs non-atomic tests
- 5:43Test data prep
- 10:16JavaScript in automation
- 13:10Avoiding duplicate coverage
- 16:46Logging without UI
- 17:30Project data via API
- 19:19Atomic tests in parallel
- 21:05Q&A
- 22:55Slow test environments
- 23:37Test scope in E2E phases
- 24:44Reducing test flakiness
- 25:47Closing remarks
related talks
38mThe Power of Atomic Automated Tests
23mAdaptive Testing
09mTest Time Execution! Why it Can't be Ignored?
41mFunctional Tests (Selenium) in Practice
21mScaling Test Automation: Building Reusability for 10x Speed and Efficiency
43mHyper-Parallelism - The Panacea of Automated Testing?
01h 14mContinuous Testing
24mSupercharging CI: Test Suite Optimization for Speed, Stability, and Savings