- Make sure your hardware is appropriately sized. TRIRIGA will typically help you with this. You just need to provide them the details of how you use (or anticipate to use) the system, such as number of users, average number of new objects you add per day, number of objects in the system etc.
- Verify that your database and application/process servers have sufficient memory allocated to them. Make sure you have sufficient memory on your severs to support the values you use
- Optimize portal performance by removing as many queries as possible from portals and eliminate or at least cut back on the use of auto-refresh.
- Upgrade to a current platform release – TRIRIGA is constantly improving the platform. Check the release notes for all of the more current platforms to determine if any new enhancements can help you.
- Try to determine if the performance issues are related to the database or to process/app servers. Check memory & CPU usage to help determine this
- Have a DBA capture long running SQL queries by running Statspack, and try to determine where they originate in the system. Many come from queries within the system that can be tuned to perform better. I have a post on my blog about this: http://www.tririga.info/2008
- Disable any unnecessary workflows. For example, if you don't use out of box approvals or notifications you can bypass a lot of system overhead by disabling these workflows, even on a per-module basis. (Be careful here if you are not familiar with workflows)
- Check the database for disproportionate table sizes. I had a customer with 250 million rows in IBS_SPEC_ASSIGNMENTS and only 10 million rows in IBS_SPEC
- Get a count of all objects in the database to be sure what you have is what you expect; you may find yourself with 250k discussion threads that you never use. Here is an article in my blog on how to do this: http://www.tririga.info/2008/01/get-count-of-all-objects-in-trirga-9.html
- Try to isolate specific actions that perform poorly, like creating a particular type of record. Analyze what the system does internally when these actions occur. Try to determine if a specific workflow or query is causing the problem and work to fix it.
If you have anything to add to this list, please let email me and let me know.