How we reconstructed our releases at Trivago
summary
Trivago rebuilds release quality by putting QA in charge of end‑to‑end tests, running ~300 Cucumber scenarios with a homegrown troopy Java framework using Selenium 4 alpha, BrowserStack for devices and Solenoid in a 500+ session internal grid, and a cucable preprocessor to split scenarios for parallel runs. Overnight monitoring and 100x qualification catches flakiness, test suites drop from about an hour to 3 minutes for key runs, and QA literally presses the release button after passing CI.
AI-written summary · about
chapters (13)
- 0:00Introduction and speaker intro
- 1:28About Trivago and the speaker
- 5:26Old release process problems
- 8:43Central test automation team issues
- 12:42New testing processes and roles
- 15:13Troopy test framework overview
- 20:51Test grids and parallelization
- 24:30Running tests via Maven and Make
- 30:07CI/CD pipeline integration
- 32:37Test reporting with Cucumber Reporter
- 35:09Full test flow walkthrough
- 37:38Key learnings and summary
- 40:48Q&A session begins
related talks
39mBuilding The Blocks of Trust In Automation
23mAre you releasing the right product?
36mFrom Sledgehammer to Fine Brush for QA
43mLean QA: Automation, Monitoring, Experiments
28mBeyond A/B: Test automation approaches for web variants from A to Z
46mMaking Security Agile
53mWhat Happens When Everyone is QA? And I Mean Everyone!
38mScaling Automated Testing & Architecture from Startup to Enterprise