WebSphere Enterprise Service Bus, Version 6.2.0 Operating Systems: AIX, HP-UX, i5/OS, Linux, Solaris, Windows


What do you monitor?

You can monitor service component events in WebSphere® ESB by selecting certain points that a service component event reaches during processing. Each service component defines these event points, which generate (or "fire") an event when the application processes at that given point. You can also monitor performance statistics for service component events.

Regardless of the type of monitoring you intend to perform on your service components (problem determination, performance tuning, or process monitoring), you monitor a certain point that is reached during these components processing. This point is referred to as an event point, and it is these points that you select to be monitored. Each event point encapsulates the service component kind tag, an optional element kind (which are specific functions of a service component type), and the nature of the event. All of these factors will determine the type of event generated by monitoring.

Event natures describe the situations required to generate events during the processing of service components. These natures are essentially key points in the logic structure of a service component that you select to be monitored. The most common natures for service component events are ENTRY, EXIT, and FAILURE, but there are many other natures depending on the particular component and element. Whenever an application containing the specified service component is subsequently invoked, an event is fired every time the processing of a service component crosses the points corresponding to the event nature.

As an example of how events are defined for a service component kind, the MAP service component kind can directly fire events with natures of ENTRY, EXIT, and FAILURE. It also includes an element kind, called Transformation, which defines a specific type of functionality within the MAP component kind. This element also fires events with ENTRY, EXIT, and FAILURE natures. Consequently, the MAP service component kind can fire up to six different events depending on the combination of elements and natures that you specify. The list of all service components, their elements, and their event natures is contained in the event catalog.

Monitoring is a separate layer of functionality that lies atop the processing of your applications, and does not interfere with the processing of your service components. Monitoring is concerned with service component processing only insofar as it detects activity at a specified event point. When this happens, an event is fired by monitoring, which determines where the event is sent, and what data is contained in that event, based on the type of monitoring you are performing, as detailed below:

Performance metrics
If you are monitoring a service component in order to gather performance metrics, light weight events are fired to the Performance Monitoring Infrastructure. You can select for monitoring one or more of the three performance statistics generated for server-specific server components:
  • A counter for each EXIT event nature – this counts successful computations.
  • A counter for each FAILURE event nature – this counts failed computations
  • The processing duration calculated between corresponding ENTRY and EXIT events (synchronous computations only).
You can also monitor the performance of applications at the Service Component Architecture (SCA) level by using Application Response Measurement (ARM) statistics. These measures allow you to monitor an application at a much finer level of detail within the application than is otherwise available in other service component events. You can use these statistics to monitor many different points between initial application call invocations and service responses, when they use the SCA.
Service component events with business objects
If you want to capture the data from events fired by monitoring at specified event points in service component, then you would configure the server to generate the event and its data to be encoded in Common Base Event formats. You can specify the level of detail of business object data to capture in each service component event. You can publish these events to either a logger or to the Common Event Infrastructure (CEI) bus, which directs the output to a specially configured CEI server database.

concept Concept topic

Terms of use | Feedback


Timestamp icon Last updated: 21 June 2010


http://publib.boulder.ibm.com/infocenter/dmndhelp/v6r2mx/topic//com.ibm.websphere.wesb620.doc/doc/cmon_whatdo.html
Copyright IBM Corporation 2005, 2010. All Rights Reserved.
This information center is powered by Eclipse technology (http://www.eclipse.org).