quote:Originally posted by JoakimI´ve moved to a 2Ghz machine with 640MB RAM monitoring only 100 elements right now.
Diagnostics suggestions:
There are a few things you can check - memory has already been suggested, and this is key. Depending on what else you are running off SQL, you can limit the amount of memory that SQL uses to prevent it sucking up the entire stack. Another perfomance checkpoint is disks - do you have your disks in the recommended config for SQL enterprise (3 physical, 1=system/programs, 2=SQL logs, 3=SQL DBs)? Something as simple as installing an extra HDD & moving the pagefile onto it might help. Use perfmon to check your disk queues.
another thing that made a huge difference for me:I wrote a batch file to delete the temporary files on the web server(s) and used scheduler to run it every 2 hours. I have a very deep map with lots of objects (16 6500s fully populated with every port on a map page as an object). my response times are great...SQL response time is VERY important as well.see attached and change .txt to .bat to run.