Scaling based on Windows Azure CPU time?

Question

Microsoft Azure's Web Sites reports it's CPU consumption as CPU time and it's measured in milliseconds. I understand it's the time taken to run code but it doesn't help make scaling decision. We are on the "Standard" plan which gives us the entire VM (no sharing).

On the other hand, Azure Cloud Services reports its CPU consumption as "CPU %" which clearly communicates total capacity as well as current consumption to easily scale (eg: over 80% => scale up. Under 15% => scale down)

So the question is, given a somewhat slippery measure like CPU time without a measure of capacity, how do you determine what's the residual capacity to make scale up/scale down decisions for azure websites?

Answer 1

I just took a look at my Azure websites setup, and it appears to me that while the monitoring metrics only give us CPU time, the automatic scaling is setup by targeting CPU percentage.