It was quite some time ago so I don't remember exactly, but I've basically used one of the usual benchmarks that everyone uses to measure the number of requests per second that a server can sustain.

(I've rented another VPS in the same data center and ran the benchmark from there.)