Create machine(s) for running Tor non-exit relays for network health work

While we have plenty of public data available for estimating the health of the whole Tor network and particular relays, it is necessary to run non-exit relays, too, to both understand and fix (overload) issues e.g. caused by DDoS attacks against onion services and Tor bugs that affect the health of our network.

Thus, we'd like to have a machine (or maybe we need two different ones) where we can run two non-exit relays on. @dgoulet and @gk should have access to those (or maybe something like network-health group where both of us belong to might make more sense, dunno).

@dgoulet will provide some specs, so we can get that task actually going. :)