Configuring monitoring for data capture

Data capture is based on a publish/subscribe model. To specify the data that you want to record to a database, you configure your message flows to emit event messages. These messages contain the data that you want to record. The messages are published to a topic that you specify in a DataCaptureSource configurable service. This configurable service also contains the name of the DataCaptureStore configurable service that represents the database to which the data is to be recorded. The broker subscribes to the topic and routes the published messages to the database.

The mechanism that is used for configuring message flows to emit event messages for data capture is also used for business-level monitoring. This mechanism is summarized in the concept topic Business-level monitoring, and is described further on in this topic. Because data capture and business-level monitoring can share event sources and monitoring events, you must ensure that the configuration for one does not inadvertently affect the other. Before you begin to configure monitoring for data capture, you need to be aware of some specific considerations that relate to the use of the monitoring mechanism for capturing data for record and replay.

• Typically, the type of data that you want to capture for record and replay differs from the type of data that is useful for business monitoring. When you are capturing data for record and replay, you probably want to capture part or all of the message bitstream data. However, if you are monitoring transactions or business processes, you might want to include only particular message fields that represent significant thresholds. See Monitoring scenarios. The requirements for data capture and for business-level monitoring might therefore conflict with one another.
• When you view your recorded data in the web administration Data Viewer, you might see one or more errors in the Exception column. You can retrieve the exception data for these errors only if you included the exception list ($ExceptionList) when you created your monitoring events. However, if the monitoring events are exported in order to create a monitoring model for WebSphere Business Monitor, the inclusion of the exception list results in a warning. See Creating a monitor model for WebSphere Business Monitor V7 or later.
• You can record messages from many different providers and sources. However, you can replay messages to WebSphere MQ destinations only.

  When you configure a monitoring event, you can choose to include all of the message bitstream data, the headers only, or the message body only. When a broker replays a message, it needs to include a WebSphere MQ message header. It handles this requirement in one of the following ways:

  • If the captured data includes a WebSphere MQ header, the broker uses the existing WebSphere MQ message header.
  • If the captured data does not include a WebSphere MQ header, the broker generates a WebSphere MQ header. All other headers are appended to the message payload.