Talk:Perfomatic:Requirements

Not sure if this really wants to be part of this project, but it feels to me like it's far too easy for humans to miss changes, even with the existence of graphserver. I think possibly one of the most valuable pieces of work that we could do around performance would be some sort of automated monitoring of all the numbers we track with automatic emailing and other notification if specific thresholds are crossed (kinda like Nagios). --dmose