Let's add some benchmarks that will let us more sufficiently understand local performance.
Specifically, let's add benchmarks at two levels:
plugin level - communicating over RPC to a plugin; let's see how this performs and attempt to get an idea as to how many requests can be multiplexed per conncetion
server level - run the server and hit it locally from goroutines to get an idea as to how many requests we can handle and what the performance looks like
A few ideas to consider later on:
should we pool the Modifier calls?
what sort of lock contention do we have in the proxier code?
Let's add some benchmarks that will let us more sufficiently understand local performance.
Specifically, let's add benchmarks at two levels:
A few ideas to consider later on: