Overall test metrics

Understanding which metrics to examine goes together with instrumenting those metrics with monitoring and then validating if the customer's and system's nonfunctional requirements are met once repeatable tests are executed. Some of the most obvious metrics to be examined are given in the following table. This table comes from a great IBM paper entitled Performance Implications of Cloud Computing by Lydia Duijvestijn, Avin Fernandes, Pamela Isom, Dave Jewell, Martin Jowett, Elisabeth Stahl, and Todd R Stockslager. It serves as a starting point for overall metrics, definitions, and respective units of measurement that can be applied to cloud-enabled portal applications or traditionally deployed portal applications and other systems connected to portal applications:

Service level	KPI	Definition	Unit of measurement
category
Availability	Service window	Time window within which KPIs are measured	Time range
	Service/system availability	Percentage of time for which...