What a staged rollout should reveal about ANRs
Application Not Responding events often trail crash reports by a few hours of attention. They feel softer — the app did not die — yet frozen main threads drive uninstalls just as quickly. During a staged rollout, watch ANR rates on the same cadence you watch fatal crashes.
Give each stage enough time for your heaviest daily paths to run: login, primary content load, checkout or booking if you have one. A two-hour stage on a quiet morning will not stress background workers that only fire after lunchtime syncs.
If ANRs climb while crashes stay flat, look at newly added blocking work on the main thread — image decoding, synchronous preference reads, or network callbacks that used to run elsewhere. Pair the ANR timeline with the commit list for that build rather than hunting the entire codebase.
Release Stability Check engagements at NetApp Solutions treat ANR movement as a first-class signal during staged windows, not an afterthought in the final summary.