Creating monitor models

To create a monitor model in the Monitor Model editor, you first create a business monitoring project and then create a monitor model. The monitor model contains a monitor details model, key performance indicator (KPI) model, dimensional model, visual model, and event model.

1. In the Business Monitoring perspective, create a business monitoring project in the Project Explorer view. See Creating business monitoring projects.

   Select other projects to reference from this project as a source of event definitions. Event definitions can be in XML Schema Definition (XSD), Web Services Description Language (WSDL), or Common Base Event files.

   If you are monitoring an event source other than a Process Server application , first import the event definitions that describe the events that are emitted by the runtime system to be monitored. Once those event definitions are available, you can create a monitor model with inbound events for each of these definitions and further refine your monitor model to extract the relevant information from the events as they arrive at run time.

2. Create or import a monitor model (.mm file). The new monitor model is shown in the Project Explorer view, along with the imported event definitions.

   Integration Designer can be used to generate a monitor model with low-level information such as event definitions, inbound events, and correlation expressions.

   See Generating monitor models

3. Define the monitor details model as follows:
   1. Define monitoring contexts. Decide on the number and structure of the monitoring contexts required. In general, start with one monitoring context and create more as the need arises during modeling. See Defining monitoring contexts.
   2. Define inbound events and associate them with the event parts or event definitions that you added in step 1. See Defining inbound events for monitoring contexts.
   3. Define keys for the monitoring contexts. Determine the piece or pieces of information that characterize the monitored entity that a monitoring context is associated with, and define its keys so that they can hold these identifiers. For example, a monitoring context for an approval process might be characterized by the workflow engine's process instance identifier and a monitoring context for order tracking might be characterized by an order number. See Defining keys.
   4. To identify events of interest and associate them with the correct monitoring context instances, complete the filter conditions and correlation expressions for the inbound events. See Defining inbound events for monitoring contexts.
   5. To detect any special conditions relevant to the business domain, define triggers. You can specify when the trigger condition is evaluated. The trigger will fire when the condition evaluates to true.
   6. To process information from the events, define metrics, counters, and stopwatches. Define when counters and stopwatches are updated, and add the expressions used to calculate metric values. See Defining metrics.
   7. Define outbound events if required. Outbound events can be received by any event-processing application, or an outbound event from one monitoring context can be an inbound event to another monitoring context or to a KPI context. Outbound events can also be received by the Monitor action services to generate alerts, or you might find it more convenient to create alerts directly in the dashboards. Based on each trigger that can cause the outbound event to be sent, specify the values to assign to each event attribute. See Defining outbound events for monitoring contexts.
4. Define the KPI model by creating a KPI context and then creating KPI definitions. Specify the initial targets and ranges to affect the display options in the dashboard. Select whether each KPI is based on a metric and aggregation function or on a calculation based on other KPIs. See Defining or creating key performance indicators (KPIs).
5. If you plan to perform dimensional analysis, define the dimensional model to aggregate information across monitoring context instances. See Defining dimensional models.