2. Sensing (Monitoring) the Service State : 2.1 Monitoring Approaches : 2.1.4 Monitoring Criteria
   
2.1.4 Monitoring Criteria
When monitoring the delivered service to understand its current behavior, the need to scale out or scale back can be done across several main categories. Each category has its own benefits and drawbacks. The ideal system considers metrics from multiple sources to make the best decisions regarding how to adapt the system to provide the preferred service level for the end user. The following table describes each category of monitoring criteria.
Table 1. Monitoring Criteria Categories
Category
Description
Example
Infrastructure
Utilization of specific infrastructure resources such as the following:
*CPU or memory utilization.
*Disk latency and bandwidth.
*Any metric that describes the health or utilization of the infrastructure.
CPU utilization greater than 80% on web tier virtual machines.
Application
Consumption of application-specific resources such as active sessions.
Active sessions per web server > 200.
End User (Real)
The response time of a live user transaction exceeds acceptable levels. Measured from the perspective of real users. Can be real time.
*Load time on a specific object > 250ms
*Page latency > 100ms.
End User (Synthetic)
The response time of a synthetic user transaction. Measured by executing synthetic transactions against application.
*Load time on a specific object > 250ms.
*Complete synthetic session > 5s.