Profiling in native and React Native

Profiling collects data while someone uses an app, helping teams find performance problems and check app health. The guests describe bottlenecks, growing memory use, and battery drain as issues to investigate. They stress that measurements need context: high CPU use may be expected for some tasks, but can be a warning when ordinary interactions cause it.

The discussion covers response and render times, frame drops, memory, CPU, network, and user engagement. At 60 FPS, each frame has about 16 milliseconds to render; missed frames can make an animation appear to skip. React DevTools Profiler and Hermes Profiler help trace rendering and JavaScript work, while Why Did You Render, React Native Performance Monitor, Flashlight, and Reassure address repeated renders, FPS, release builds, or performance regressions.

For Android, the guests recommend Android Studio Profiler for CPU, memory, and thread data, plus system traces, Logcat, Macrobenchmark, and Microbenchmark. For iOS, they describe Xcode’s overview and Instruments for examining CPU, memory, network, and other activity. They note that the tools differ, but the underlying metrics and reasoning transfer across platforms.

The guests suggest using user feedback, QA reports, and production user flows to identify screens that may be slow or lose users. From there, profile important paths, investigate suspicious metrics, fix issues, and keep checking for regressions with tools such as Reassure or Android benchmarks. Their approaches differ: one starts with Flipper or a familiar profiler, while another starts with a broad native overview and follows clues.