Was looking at some data in the interface tonight and I realized something. In Scatter Charts where there is an error, the average times where far higher than they should be.

Guess what?

I was trying to be cute by combining the average time calculation and the availability calculation into a single query. But by doing this, I was including error measurements in the calculation of the average component and total times.

Not good.

Fixed now.

UGH!

Related Posts

• mon.itor.us Outage
• State of the GrabPERF Update — August 2007
• GrabPERF: Content! Watch your content!
• GrabPERF: Some System Statistics
• GrabPERF: Measurement Cleanup