KPI instances and KPIs

You can create a key performance indicator (KPI) for each type of data that the Web Service Data Collector collects. Optimize creates and stores data that corresponds to the KPI in a KPI instance. For example, for a KPI that tracks Broker client connection state, Optimize creates one KPI instance for each tracked Broker. For more information about the Web Service Data Collector, see Web Service Data Collector.

When you create a KPI, you can specify a collection interval to summarize the data stored for the KPI instance associated with the KPI. The Analytic Engine runs algorithms on the data to produce a variety of diagnoses. These diagnoses enable Optimize to automatically and continually compare summarized data samples to a baseline of historical behavior and indicate when types of data are deviating from that baseline.

Note:
You can provide your own definitions of when types of data are out of compliance by defining rules. For more information about defining rules, see Chapter 5 “Defining Rules” in webMethods Optimize User’s Guide.

System KPI instances store data about applications, and business KPI instances store data about business data and processes. In the My webMethods interface, each type of data that Optimize collects is represented by the KPI instance in which the data is stored.

KPI instance	Function
Sum	Sums the data points collected over each collection interval. An example use for this type of KPI instance is to tally process step errors.
Average	Calculates the average value of the data collected over each collection interval. An example use for this type of KPI instance is to track CPU usage.
Last Value	Stores the value of the last data collected in the collection interval. An example use for this type of KPI instance is to track the size of a queue.
State	Stores the value of last data collected in each collection interval, where the possible values are 1 (online) or 0 (offline). An example use for this type of KPI instance is to track the state of an application.
Count	Counts the number of data points over each collection interval. An example use for this type of KPI instance is to count the number of older events.