Home of the SQLSaurus
The tongue-in-cheek response:1) End users writing poor queries2) Power users writing poor queries3) Junior developers writing poor queries4) Senior developers writing poor queries5) Software architects writing poor queries.
You know, I started with that, but I was having trouble filling a page ;)
Another quick note on client timeouts. The knee jerk reaction is very frequently "can we increase the timeout". The answer is no. That query is performing poorly. It is a big ugly troll eating up time, and the answer is not to feed the troll more time. You want to take that troll to the gym, and get him running faster by tuning him instead!
You bring up an excellent point about the regular system process histories, like those in the SQL Agent. I've also seen that happen with job history and job step history in MSDB where that database grows to enormous size too.Great post, btw!-Kevin