I hear what you’re saying about windows, but when I wrote collectl the actual focus was on High Performance Computing and at the time that meant linux. Believe it or not I did fiddle around with some of the windows performance counters and even hacked up something that would read the counters and then call the collectl print routines. And it sort of worked, but there are just so many existing tools for windows I figured we didn’t need another and I didn’t want to distract myself with dealing with yet a whole other operating system.

A personal philosophy I’ve tried to live by with collectl is to ‘do a couple of things real well rather than a lot of things that are just mediocre’ and so there are no GUIs for collectl and no centralized cluster monitoring. Collectl’s focus is on gathering/logging metrics efficiently and making them available for others to consume via a plugin mechanism and by making them available over a socket.

-mark