This section tells you about monitoring and improving the
overall performance of the service components in your enterprise service bus. This section also
helps you when troubleshooting problems.
About this task
The facilities for monitoring the WebSphere® ESB are
provided by the underlying WebSphere Application Server. The techniques
used are the same. This information provides an overview of monitoring
the WebSphere ESB.
It is supplemental to the monitoring information provided for WebSphere Application Server.
WebSphere ESB performance
is affected by tunable settings provided by WebSphere Application Server. Use these
settings to make adjustments which improve the runtime performance
of your application. Many service applications run successfully without
any changes to the default values of the tuning parameters. Other
applications might need changes to achieve optimal performance, for
example by setting a larger heap size.
Tuning tips and best
practices for WebSphere ESB are
also discussed in the WebSphere Business Integration
V6.0.2 Performance Tuning Redpaper, available at http://www.redbooks.ibm.com/redpieces/abstracts/redp4304.html.
This IBM® Redpaper provides performance tuning tips
and best practices based on the performance teams experience for WebSphere Process Server, WebSphere ESB, WebSphere Integration Developer, WebSphere Adapters, and WebSphere Business
Monitor.
- Monitoring the overall performance of the system
It
is essential to monitor and understand the performance of every system
involved in running your WebSphere ESB.
This includes application servers, databases, enterprise information
systems used for service providers, and any other systems critical
to the WebSphere ESB.
If any system has a problem, it might have an effect on other systems,
causing the throughput of requests through the ESB to slow. IBM and
several other business partners use the WebSphere APIs
to capture performance data. This data can be incorporated into an
overall 24-by-7 monitoring solution. WebSphere ESB uses
the Performance Monitoring Infrastructure (PMI) data
available in WebSphere Application Server to
help monitor the overall system health. PMI provides data about system
resources, application resources, and system metrics. Many statistics
are available in WebSphere Application Server. Understanding
the statistics that most directly measure your site can help you detect
and correct problems.
You can use the Tivoli® Performance
Viewer to start and stop performance monitoring; view PMI data in
chart or table form as it occurs on your system; and, optionally,
log the data to a file for later review.
- Monitoring the throughput of SCA requests
This
information helps you understand the flow of requests for specific
mediation modules, and for specific requesters and providers. This
detail is important for an in-depth understanding of which system
component is using specific resources. Typically at this stage, you
deploy some type of trace through the service application, or thread
analysis under load condition techniques to isolate areas of the application
and particular interactions with the back-end systems that are especially
slow under load. In this case, WebSphere ESB provides request
metrics to help trace each individual transaction as it flows
through the ESB, recording the response time at different stages of
the transaction flow. In addition, several IBM development
and monitoring tools that are based on the request metrics technology
(for example, Tivoli Monitoring for Transaction Performance)
are available to help view the transaction flow.
For each SCA
module deployed on WebSphere ESB, requests being
processed are held on queue points and in the data store for messaging
engines. You can display the data for SCA requests, and if appropriate
take further action to manage the throughput of SCA requests.
- Monitoring Common Base Events fired at points
in the processing of service components.
You can configure WebSphere ESB to
capture the data in a service component at a certain event point.
The Common Event Infrastructure (CEI) is used to provide basic management
services for events. The format of those events is defined by the
Common Base Event specification. You can have this event data published
to the logging facilities, or you can use the more versatile monitoring
capabilities of a Common Event Infrastructure server.