You can create three types of rules: KPI rules, event rules, and threshold rules.

KPI rules evaluate a KPI’s aggregated data at the KPI’s collection interval. For example, suppose you create a KPI that sums order amounts on a 5 minute interval. You could then create a rule that sends an alert when the summed order amount exceeds a certain value. For more information about creating KPIs, see Administering webMethods Optimize.

Event rules evaluate data points as they are collected rather than waiting for a KPI collection interval. Optimize maintains state for event rules so that you can track data over multiple event intervals (for example, you can define a rule that alerts you when a webMethods Broker is down for three event intervals).

Threshold rules evaluate a single collected data point when it is collected (for example, queue size greater than 1000 or order amount less than $500). Optimize does not maintain state for threshold rules.

KPI and Event rules differ from threshold rules in that they maintain a state. When activated, they stay in violation until their expression condition returns to compliance. These expression conditions can combine attributes from multiple KPIs and event types. Note that Event rules don't actually operate on a KPI, but they are artificially tied to a KPI in a definition. This doesn't affect rule functionality, however.

Threshold rules, on the other hand, do not maintain state and cannot use multiple event types in their expression. A threshold rule operates on a single event identified by an event map, whereas KPI and event rules operate on one or many KPIs. Threshold rules are activated whenever the threshold condition defined by the rule expression is violated. The sticky function is not available for threshold rules because they do not maintain state. In addition, threshold rules do not distinguish between event data dimensions. Finally, one or more of the expression arguments in a threshold rule can be an event with transaction attributes.