Results of perf_tests/soak.py run by a remote tester against the staging service (guide: docs/soak-test.md); analysis in docs/performance-report.md.
summary.md: the run’s own summary (verdict PASS).windows.csv: one row per one-minute window (1,440 rows).meta.json: the run’s settings and the tester’s platform.windows.jsonl (the same rows as JSON) and canaries.jsonl (7,200 known-answer checks, all correct); they are kept by the maintainers.The tester’s label was replaced with remote-tester in these copies; nothing else was changed.