You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
I'm already dogfooding Bencher with Bencher, but all of the existing benchmark harness adapters are for micro-benchmarking harnesses. Most HTTP harnesses are really load testing harnesses, and load testing is different than benchmarking. Further, I'm not looking to expand Bencher into load testing anytime soon. That is a very different use case that would require very different design considerations.
Instead, implement a micro-benchmark that runs against the Perf API endpoint and dogfood the results with Bencher. This will require a sizable test database to make sure that performance regressions like #363 are caught in CI.
Track the performance of the Bencher Perf API with Bencher.
I'm already dogfooding Bencher with Bencher, but all of the existing benchmark harness adapters are for micro-benchmarking harnesses. Most HTTP harnesses are really load testing harnesses, and load testing is different than benchmarking. Further, I'm not looking to expand Bencher into load testing anytime soon. That is a very different use case that would require very different design considerations.
Instead, implement a micro-benchmark that runs against the Perf API endpoint and dogfood the results with Bencher. This will require a sizable test database to make sure that performance regressions like #363 are caught in CI.
Read more: Why SQLite Performance Tuning made Bencher 1200x Faster
The text was updated successfully, but these errors were encountered: