Microsoft`s definition of 「All Instance」 for calculating average percentage of CPU usage is a mystery for me !!!.

Question

Hi, I was looking up in the 「How to Auto Scale」 document
 　http://azure.microsoft.com/en-us/documentation/articles/cloud-services-how-to-scale/
and got stuck in the below paragraph, the part that describes how to calculate the
 average percentage of CPU usage. 
==========
All instances are included when calculating the average percentage of CPU usage and the average is
based on use over the previous hour. Depending on the number of instances that your application is using, it can take longer than the specified wait time for the scale action to occur if the wait time is set very low. The minimum time between scaling actions
is five minutes. Scaling actions cannot occur if any of the instances are in a transitioning state.
==========
Does any one know the microsoft`s definition of 「All Instance」 ?
Whether the virtual machines that are shut down but has been set into auto scaling group count as one of them?
The question is that the response time, the data shows up in the management portal delays a lot comparing to the
real time. I was wondering if the definition of the calculation described above is truly what the user`s want or not.
I`m　all ears to any information. Thanks in advance !! 

Answer 1

Hello Tomo Shimazu,
Instance is a presence of a service eg :- VM is an instance. Copy of VM is another instance which is 2 instances of VM.
Although Microsoft’s pricing model is readily available and even offers a simple calculator (http://www.windowsazure.com/en-us/pricing/calculator/advanced/) to assist
in the process, it may be difficult to know how much of each resource you will need to get an estimate.
To understand on calculating the compute instances and the VM size, let’s first take a look at the VM role sizes offered by Microsoft.
The prices per month for each, at the time of writing, are $30, $90, $180, $360, $720 from ExtraSmall to ExtraLarge respectively. So with the exception of the
transition from ExtraSmall to Small, going to the next size VM is exactly twice the cost, which, is mirrored in the increase of resources you get at each level. With each step, CPU Cores, Memory, Disk Space, and Bandwidth are all doubled from the previous.
If this is the case, then, is there any advantage to any one of these over the others? The answer is yes. In the majority of cases, the best bet will be to go with the small instance, with the reason being that because all of these resources scale equally
to cost it is possible to achieve the exact equivalent to a larger VM simply by increasing the number of instances. For example 8 small instances is equivalent to one ExtraLarge instance, with the advantage of the fact that when not needed these small instances
can be turned off and will cost nothing.
By hosting the application in this manner, it increases the effectiveness of the number one reason that a business would transition to the cloud anyway – Scalability.
Obviously 16 smaller instances can be more finely adjusted to the application’s usage than 2 ExtraLarge can. If traffic gets higher or lower than expected, two Small Instances can be added at $180/mo for the time that they are running, versus adding another
ExtraLarge for $720/mo.
The only exception to using the smallest instance, is in the case of the ExtraSmall, which offers only 1/20 of the bandwidth of the Small, making it only feasible
for very lightly accessed applications or, more likely, a QA environment.
From this point, it becomes much easier to estimate the compute instance requirements of migrating the application to Windows Azure. Take the current server(s)
total resources (CPU, RAM, etc) and find how many Small instances it would take to recreate it. This gives a good starting point, however, remember that these instances can be turned on and off to meet demand, and beyond the first instance, there is no charge
when the instances are off. This can lead to significant cost savings, and is the primary incentive to migrating to the cloud.
For azure pricing information, refer
pricing details.
If you still unclear about pricing and billing, you may raise a service request with billing team
here
Hope this helps.
Regards,
Shirisha Paderu