I've been tasked with gathering performance metrics on a few database servers that have one TempDB data file. It seems prevailing wisdom is to have one TempDB data file per core up to a certain point. Both servers in question have 4 CPUs, so we're going to add three more on each.
My question is which performance metrics should I gather to get a solid depiction of where my servers are at? I want to gather data pre-change and post-change for comparison.
