Mobile Application Performance Done The Right Way
summary
Claims that combining tracing and APM finally brings server-grade observability to mobile and shows concrete fixes: a visual trace tool (demoed via helpfulprofiler.com) exposes a Kotlin coroutine dispatched in series due to a misplaced scope/curly brace and a 600 ms camera hangup delay in Signal that drops ~43% of that flow time and, when removed, cuts the hangup-to-menu latency by about 0.5 s (~68%). Recommends automatic instrumentation, avoid sampling, and focus on user flow tracing rather than manual traces.
AI-written summary · about
chapters (9)
- 0:00Introduction and speaker background
- 3:05Why mobile app performance matters
- 7:36Problems with current mobile performance tools
- 9:20How performance is done wrong
- 11:27How to do performance right
- 14:15Live demo of performance issue
- 19:50Real-world example with Signal app
- 24:06Summary and call to action
- 27:36Q&A
related talks
21mMobile Application Performance Done The Right Way
17m“How’s the App Doing?” Bringing Mobile Into Your Reliability Picture
43mPerformance in a Kotlin World
34mClient Tracing At Scale
30mLearn from developers: Lyft and Google Maps share their experiences optimizing app performance
35mIn Search of Performance: the Subtle Art of Android Profiling
38mA Busy Android App Developers' Guide to Perfetto
39mAndroid App Performance in a Nutshell