If You Can’t Measure It, You Can’t Manage It

“No news is good news” just isn’t valid in the Operations SA world. You need to be watching stats on every little process, at all times.

Which is impossible, of course.

So you need to have automated collection of stats on all your processes, and automated alarming when those stats step outside of the normal bounds. And each time you find and fix an issue, you have the potential to need to add yet another statistic with yet another alarm.

All too often a problem is found to be caused by a process that is not monitored enough, or at all. You probably don’t have too many stats, but too few – no matter how many you are currently collecting.