2025-11-12 · Hana Mori
Measuring wake latency without breaking sterile build habits
Wake latency stories often start with a heroic trace capture and end with a configuration nobody can reproduce. We keep hooks narrow, name them like product features, and tie each marker to a single acceptance clause so verification teams know what “done” means.
When sterile build rules forbid certain macros, we prefer thin wrapper units compiled only in engineering branches. Those wrappers forward to stable production entry points, which keeps divergence visible in code review instead of hiding in linker scripts.
If your hazard analysis mentions brown-out recovery, measure wake paths under both nominal and reduced rails, but document the power supply assumptions explicitly. External reviewers care less about the absolute millisecond count and more about whether the measurement conditions match the risk narrative.
Finally, archive the harness scripts beside the trace files. Future-you will thank present-you when a silicon spin forces a partial replay six months later.