This is the userver portion of a benchmarking test suite comparing a variety of web development platforms.
This benchmarks comes in two configurations: userver and userver-bare, where both configurations use exactly the same handlers code, but userver-bare replaces default http implementation of userver with custom one.
You see, userver being feature-rich framework widely used in production comes with a lot of useful functionality built-in (metrics, dynamic configuring, logging/tracing, congestion control etc...) none of which is of any use in benchmarks; although most of that can be disabled via configs, some parts remain, and these parts aren't free.
The aim of userver-bare is to explore practical limits of lower-level userver functionality when performance is an absolute must, while still being idiomatic userver code.
http://localhost:8080/plaintext