That’s not very useful.

This is what I need to know:

[Disk Space] 14 days and 10 seconds left

If this is shows up on a critical system, I’ll begin to panic. I only have two weeks to shift resources around or beg for more storage. Given how long adding specialized storage could take because of budget concerns… I must act immediately.

I pay less attention to metrics I can’t take direct action.

Here’s another example of changing metrics to suit me. I used to measure packets/second and bandwidth for individual systems. I had colorful charts spanning days, and that was not useful at all. All I cared about is if packets/second was 0 for more than 5 minutes. This could mean I’m not recieving data. I could take immediate action by looking at that node in the system. Bandwidth was a nice to know and we never exceed the capacity of our bandwidth.

What is more important is knowing if I dropped any data.