Previously it focused on per-request timing. This update adds rolling-window aggregation to detect performance trends.

Gathering the stats in-memory is neat - I'd expected this would be writing to a DB but that's not how it works: https://github.com/yeongbin05/django-xbench/blob/f63316126b5...

Terrific project. I use Strawberry GraphQL so about all of my endpoints will be /graphql unfortunately. It would be nice if it could parse off a header to qualify which bucket it goes to. That way I could send over the operation and django-xbench takes care of it.

I see you are going have N+1s on the roadmap. I currently use django-zeal which does a fantastic job detecting them but there isn't a visualization. I'm keenly watching where you go with this.

I need more django perf in my life.