Glad to see Hanami still kicking, but this kind of benchmark is not super useful. The vast majority of time spent in Ruby web apps is on business logic and not web framework code (except for Rails, maybe). So at the end of the day, running your custom, heavy code on any framework is likely to yield the same result because >90% of your time will be in your code.
Hm, in API use, your framework can actually add substantial overhead, especially if you rely on caching.
I haven’t been writing a Ruby framework for a while, but quite some time in the request/response cycle is spent in the security stack (cookie encryption/parsing, CORS mitigations, etc.) and having frameworks that provide faster API throughput by having some tuned stack helps. Comparing this stack to others is certainly not without use.
I was surprised Sinatra did so poorly on the benchmark. Does anyone here know why?
Was thinking the same! Hoping someone has some pointers, I’ll probably do some digging later to find out why
This is exciting. I was looking for something like this just last week.