It’s funny how when you’re troubleshooting a performance issue on your servers that suddenly made the load average spike to 14 (350% with four CPU cores) at 6:30am on a Sunday morning (yay!) all the stats look like garbage until you figure out what it is and then it’s so glaringly obvious that you spend the rest of the day kicking yourself for not seeing it immediately.

Note to self: Next time make coffee, drink coffee, and only then log on to munin to troubleshoot.