diff --git a/Testing-Guide.md b/Testing-Guide.md index 6c8d4d2..20a756d 100644 --- a/Testing-Guide.md +++ b/Testing-Guide.md @@ -390,9 +390,9 @@ stay up to date with it: | https://localhost:8444/tu/{id} | #158 | Trusted User Proposal | Trusted User | | https://localhost:8444/addvote | #159 | Trusted User Proposal Create | Trusted User | -**Status Legend:**
-- Incomplete: Needs testing. -- Complete: All done. +Each item has its own issue, linked in the **Status** column. Feedback can be provided in the form of comments to these issues, or through their own issue using the Feedback template. + +In any case, take a look at the Feedback template and use the same template for comments you may include on one of the `[User Testing]` issues. #### Development Branches @@ -452,11 +452,18 @@ One of our sub-goals of this port is to increase performance and efficiency overall across the service. Lower benchmark times indicate that code is performing better overall. +When performing benchmark comparisons between PHP and FastAPI, users should opt +into defining `FASTAPI_BACKEND=gunicorn FASTAPI_WORKERS=5` as they `docker-compose up -d nginx`. Our `php-fpm` service uses `5` workers by default, and so using `gunicorn` with `5` workers does the best we can to match core usage with PHP. + To perform benchmarks, the [hyperfine](https://github.com/sharkdp/hyperfine) -tool can be utilized. +tool can be utilized: $ hyperfine --warmup 20 --min-runs 500 'curl https://localhost:8444' +Or, the [oha](https://github.com/hatoo/oha) tool: + + $ oha -c 8 -z 1min 'https://localhost:8444' + These kinds of benchmarks matter more on particular routes with specific behaviors. We are primarily concerned with the performance of database query related tasks like package search, package/pkgbase viewing, requests