Loading…
A short voice session: one question, one follow-up, then instant feedback — up to 5 minutes.
Compare power-related activity before and after the release on representative devices. Correlate CPU, wakeups, network, sensors, and background execution with feature usage.
Segment by OS, device class, and enabled features to avoid blaming the whole release uniformly. Look for leaked location subscriptions, tight retry loops, excessive polling, or work that continues after its owner disappears. Reproduce with controlled workloads and use platform power diagnostics. Verify that the fix reduces measured activity without silently dropping required user work.