Has anyone done an performance research on the overhead the ktor clients provide over their native counterparts? (like ktor-java, vs java client directly, or ktor-okhttp vs okhttp directly)
I was doing some testing via the
client-benchmarks provided by ktor, and found close a 3.5x or so difference in max RPS between ktor wrapped clients vs the client directly.
(i.e. native clients had 3.5x more RPS than ktor wrapped client)
github.com/ragboyjr/…/docker-benchmark#… - I captured some of my findings here. would be curious if others have seen something similar?