I have my database with indexes in source control and they go through dev/qa testing before being deployed to PRD.
Any suggestions on working with queries that seem like they would benefit PRD, but do absolutely nothing in QA (or visa versa)?
PRD has roughly the same data... similar automated load/reporting processes, but the user usage is totally different in QA... as well as being virtualized with far less memory.
Optimizing fully for PRD, you deploy the changes to QA and it can be like a shackle harming performance instead of helping
Anyone have strategies for this?