Processing Units (All OM Types)

To configure a processing unit (PU), you add the items you configured earlier, and any additional properties required. If you do not find a configuration item you require, click the appropriate tab and add it, then return to Processing Units tab and continue configuration.

One processing unit named default is provided out of the box. You can change this name. It has no significance, except that TIBCO Administrator expects a processing unit of this name by default, which can be useful for testing purposes.

Processing units are referenced in the site topology file used by TIBCO BusinessEvents Monitoring and Management component (see Basic MM Configuration in TIBCO BusinessEvents Administration).

PUs with Unique Agent Instance Properties

Depending on configuration, some processing units can be deployed more than once in a cluster. Others have unique configuration details that make them deployment-specific, that is, that limit them to being deployed only once in a cluster. Configuring the following CDD settings for a processing unit makes it deployment-specific processing unit:

Agent Key
 At the processing unit node, you can associate a unique key with a selected agent class. This key identifies an agent instance uniquely at runtime. The purpose of the agent key is to retrieve scorecards from the backing store. Scorecards are local to an agent and the key enables the correct scorecard to be returned to the correct agent.
Agent Priority
 The agent priority determines which agents of a given class are active, when fault tolerance is used. Each deployed agent of an agent class can have a different priority; however, if the agent have the same priority then cluster decides which agents it has to activate.

See Deployment-Specific Processing Units and Global Variables in TIBCO BusinessEvents Administration for other ways a processing unit can be deployment-specific.