maybe add some notes on benchmarking

i think this probably belongs on nanos.org instead but might have a section on how to benchmark various things properly:

defining what you want to benchmark (throughput, latency, storage vs/networking, writes/reads, random writes/reads, large/small, etc..)
isolating components from each other (isolate {virt cost, networking config, linux<>nanos variance, app specific such as multi-process, etc.})
suggested envs/reproducible scripts
networking: usermode vs bridge; host vs guest; same class c vs pub inet;
linux guest tuning (port limit, mq, wrk vs/ab;}

nanovms / ops-documentation