<gingeropolous> status on reproduction: it claims it reproduced the 12/12, but some key metric is missing because it was using the get_peer_lists call within the simulated network, and at the high n the call essentially timed out at the peak of all the stuff. I'm getting it to re-wire the measurement mechanism to use host-local scripts, so it [... too long, see mrelay.p2pool.observer/e/scGMm6sLODBKZ0ZP ]