Give the ability to drill down by database, machine, user across a time period instead of within a time period

Selectable statistics collection phase.

We have hundreds of databases and to collect data for a longer historical period across the board would be a huge undertaking storage and network wise.
The idea has merit though as I too noticed that it is difficult to make decisions on just 24 hrs of data.

Considering what is required relates only to occasional circumstances then perhaps if we could, when required, select to collect data over a longer period and select that period e.g. few-days, week, month. This would then collect a small set of data on a single-few items that would assist our diagnostics and save on resources  e.g. network, CPU and storage.

Whilst making manual images of this week's Friday run statistics to compare with the next nothing competes with live data.

This would  be beneficial to us as well. We have 20-30 databases. I would like to make a performance improvement in one database and see the effect on the one database  before applying to all of the databases.

layfnatu2, I think I understand what you are asking for.  Let me know if you don't think I've captured it correctly. 

I would describe it as the ability to simultaneously filter and pivot wait time by multiple dimensions.  Kind of like using an Excel pivot table to query an SSAS cube.

comment add by Kathy from layfnatu2