25
RESOLVED:
The core of the problem was an httpd issue. As fair warning to anyone running as powerful a CMS as Xoops, you will need to look towards a dedicated server as you push the 2GB bandwidth / month barrier. A large client / thread ratio was the cause, where httpd would bind up to the point it could simply no longer handle requests. Coupled with the small delay in SQL, multiplied by 800,000 hits in a day, and it was no wonder we folded like an empty winerack.
I'll updated the first post to mark its resolved status, and hopefully this will help others who are having similar problems. Be weary of your server load, XOOPS will consume a lot of resources during peak traffic.
Lesson well learned.