Mobile Device Farms Demystifying Testing Strategies Between Theory and Real World Practice - Sergio
summary
Company runs Appium tests in-house because the VoIP certificate breaks cloud resigning so cloud farms cannot install a working app. Built a small lab of five iPhones (one iOS16, two iOS17, two iOS18) on Mac hardware with powered USB hubs and a “finger bot” for remote restarts. CI uses GitLab runners limited to two concurrent jobs; hooks + login add ~35s while a combined end-to-end test drops total runtime from ~30 minutes for many isolated tests to ~11 minutes. Team reboots after each feature file and retries flaky backend failures.
AI-written summary · about
chapters (8)
- 0:00Introduction and overview
- 1:46In-house vs. cloud solutions
- 3:48Comparative analysis
- 11:04VoIP and technical constraints
- 13:57Overcoming infrastructure gaps
- 19:06Mitigating test flakiness
- 23:21Pragmatic testing strategy
- 29:44Conclusion and Q&A
related talks
39mOverview of Device Farms for Mobile Testing
14mMobile Farming for Perfection
24mAutomated Mobile E2E Testing
25mBuilding a scalable device lab with Ansible
29mPros and Cons of Device Farms Available for Android
35mDockerization of a Real Mobile Device Farm and Scalable QA Automation Ecosystem
52mBuilding an Advanced Automation Solution Based on Appium
26mMobile Test Automation Frameworks – From the Basics to Usage