benchmark

Detect webpage performance regressions by capturing baseline metrics and comparing before/after results on each PR. The benchmark tool measures page load times and Core Web Vitals, flagging any degradation introduced by code changes.

Track Core Web Vitals trends over time by using the --trend option to analyze historical performance data. This evaluates multiple pages across PR history and stores reports under .gstack/benchmark-reports for ongoing visualization.

Establish a baseline for page load times and resource sizes by running a performance audit on a target URL. Use the --baseline option to capture initial timing measurements, which subsequent pull request diffs will compare against.

Measuring webpage performance metrics requires a gstack-enabled environment with the browse daemon running. This setup allows the tool to capture timings and persist generated reports under the .gstack/benchmark-reports directory.

Compare performance diffs for multiple pages simultaneously by utilizing the pages list feature alongside the --diff option. This evaluates a defined set of pages against an established baseline to identify regressions across the application.

Performance baselines provide a historical reference point, enabling accurate regression detection rather than isolated snapshots. Comparing current metrics against a baseline on every PR ensures new code does not silently degrade page speed or resource efficiency.