Understanding real usage.
WHAT THE PLATFORM PROVIDES
Crash reports, aggregated Performance metrics from real devices Installation and usage figures Subscription and sales reporting
WHAT PERFORMANCE METRICS COVER
Launch time Hang rate Memory Disk and battery use
WHY THEY MATTER
They come from real users on real devices, which laboratory testing does not.
WHAT TO MONITOR
Crash-free rate Hang rate Launch time Any metric worsening after a release
WHAT SYMBOLS ARE FOR
Making crash reports readable.
WHAT TO ENSURE
That symbols are uploaded for every build.
WHY
Without them reports are unreadable.
WHAT TO ADD
Your own crash reporting, for faster and more detailed reports.
WHAT TO COMBINE THEM WITH
Breadcrumbs recording what the user was doing.
WHY
A stack trace without context frequently does not explain the cause.
WHAT TO MEASURE BEYOND TECHNICAL HEALTH
Whether users complete the tasks the application exists for.
WHY
Every technical measure can look healthy while nothing useful is happening.
