android-perfetto-analyze

You can analyze an Android Perfetto trace by running trace_processor SQL queries against your .perfetto-trace file. This workflow extracts concrete metrics like frame timings and thread workloads, outputting clearly labeled findings suitable for performance reports.

Identifying main-thread work and GC pauses is best done using SQL recipes on a .perfetto-trace. The workflow executes targeted trace_processor queries, iterates to narrow down slices, and delegates parsing to a sub-agent for a concise verdict on performance issues.

You do not need to manually install trace_processor to run SQL queries. The analysis workflow automatically installs and caches the trace_processor if needed, selecting an appropriate SQL recipe to execute against your provided .perfetto-trace file.

Yes, you can use SQL to measure IPC costs and frame-budget issues in Android profiling. Running trace_processor SQL recipes against a .perfetto-trace allows the workflow to iterate queries across frames, threads, and slices to extract these precise measurements.

A Perfetto trace analysis workflow generates actionable findings by running SQL queries and writing results to a text file. A Sonnet sub-agent then parses this data to deliver a small, digestible verdict with guidance on next steps for performance issues.

Using trace_processor SQL for Android performance profiling requires a provided .perfetto-trace file and depends on iterating queries to narrow down issues. Outputs are compact text findings, meaning complex visual trace exploration is outside this SQL-driven workflow's scope.