Naively, it seems like one or both of two general problems must be responsible for the population-sensitive lag:
1) There is a network bottleneck (either at the edge of SSG's network, or inside it) unable to handle the traffic from ~1k players, and/or
2) The servers are unable to respond to incoming traffic at the rate it arrives from ~1k players
This is almost tautological.
While these problems may not be easy to solve, they should be completely trivial to monitor and track on SSG's side.
To check whether #1 is happening, you compare the traffic into, and out of, each network node, in both directions.
To check whether #2 is happening, you look at the server latency (response time to incoming client requests).
These types of server-side metrics (and a lot more) should be available at their fingertips, and yet they're asking for emailed dxdiag printouts and anecdotal video clips of people lagging out?
Dagoreth (Warden) and Belechannas (Lore-master) of Arkenstone
< No Dorfs > Fighting the Dorf menace to Middle Earth since 2008