...OR you just hit the nail on the head. maybe old siege servers had logging off; new siege servers have it on; resources are overloaded and commands are throttled. maybe it's as simple as shutting off the debug mode or lowering the severity level...
???
a sysadmin should never admit this, but I have had logs set too low in severity completely fill up a hard disk before. running out of inodes is another annoying little demon.
Another method is to have a set size limit on the log files and autodelete the okder entries when the limit is hit. The problem here is that you risk missing log entries that would have helped your investigation.
Siege ,since the server move, has regressed to the condition it was back in mid 2010 when it suffered from the same, exact problems. First the severe lag starting at an interval after Siege start then, after maintenance (and a fix attempt?), back to where most everyone can not even get in Siege, instead nothing happens for a few minutes then you get kicked back to login.
Frogster did eventually fix the problems then, checking to see what they did back then should be productive.
However to continue to **** your customers off by neglecting/refusing/failing (choose one) to fix problems with a feature they passionately care about is never a good idea.
We are only a few weeks away from the two year anniversary of Siege's beta status, these current problems are not related to its beta status but to the server move. The various beta problems like guards & towers still reflecting the level limit from two years ago are discussed elsewhere.