Deciding What Event Data to Collect

Deciding what event data to collect requires that you understand the types of data the ESS delivers and how you can use this data in your event consumer application. Although it is technically possible to request all event data for all Processes in the system, it is usually unnecessary and impractical to do so. Typically, an event consumer application is interested in a subset of event data.

Note: Statistics records excluded through the API STAT command or a DTF statistics exit, or by coding the DTF STAT.EXCLUDE initialization parameter, are not available to ESS.

Following are commonly requested event data:

  • All event data for a Process or group of Processes

    Requesting all event data provides the most detailed trace of Process execution. This information is useful when testing an event consumer application. However, it is unnecessary after an application goes into production.

  • Major event data for a Process or group of Processes

    This request is the most common subset specification. For normal production work, this request gives a sufficient trace of Process execution, and minimizes overhead for ESS processing and event records by your application.

  • Subset of event data for a particular Process, set of Processes, or all Processes

    For example, an application executes a procedure when a IBM® Connect:Direct® Process terminates. In this case, it may be sufficient to register interest in PT (Process Termination) records only. As a second example, a fault reporting application needs CT (Copy Termination) statements that did not complete normally (signalling the failure of a IBM Connect:Direct file transfer request).