Has anyone experienced limitations with the number of complexity of event frames? We have heard there are issues but not experienced any as of yet.
I believe once you get over 100 million EF you might experience some pain. In general, the recommendation is about 1 GB of RAM per 1 million EF. How many EF do you have and what do you mean by complexity? What types of queries are typically done? Some queries are inherently much more efficient than others. I know in the latest AF Server, there was some significant performance improvements for EF search.
Can you give more details about the issues you have heard?
Dan is right about the approximate scale limit. It really depends on how many attributes per Event Frames, how many have captured values, and for those that don't have captured values what the data references are.
Keep in mind that if there are a lot of changes on your AF Server, such as a lot of event frames being generated, you will need to "reindex" more frequently.
If you start scaling upward, be sure that you have OSIsoft guiding you along wrt sizing of your SQLserver database. There are various statements by various people (XLS sizing sheet, 60% of SQL size, 1Gb per 1M EF's) and there is no real guidance from OSI what really applies. In particular the sizing XLS is very misleading, as this only uses a relative small sizing of the EF records into account. The more attributes are captured the more diskspace you need.
Retrieving data ...