... but these have mostly been attributed to IIS running on the server which is 2003 R2 x64. After about 3 weeks it virtually grinds to a stop, a quick iisreset and things go back to good. This has now become a regular maintenance task as a result.
One thing that we had done for us was to put the Change, Problem and Incident workflows into their own IIS application pool and leave the other bits and pieces in the original IIS pool. This made a noticeable difference but doesn't stop IIS playing up.
We have roughly 40 concurrent users at the moment with an average of 150 new Incidents logged every day, we don't use the email feature in anger just yet. We plan to expand this to about 200 Service Desk 'analysts' and allowing our 10,000 user estate to log their own tickets.
We have an off-box SQL server which is also our NS SQL server but they sit in the same instance this is 2005 SP3. I think out of our entire Altiris design SQL is probably the bottleneck.
Interestingly enough though if you run SD in IE 7 or 8 it can be frustrating and appears to 'hang'. SD in Firefox 3.6 and 4 Beta is very very quick which makes me think that it is something to do with the client-side processing. I've not spent long trying to qualify this as IE is our corporate standard so I can't advocate Firefox and get away with it :)