The are several methods that you can use to specify service
component event points for monitoring, depending on the type of monitoring
you are planning to do.
- Performance statistics
- For Performance Monitoring Infrastructure (PMI) statistics, use
the administrative console to specify the particular event points
and their associated performance measurements that you want to monitor.
After you start monitoring service component performance, the generated
statistics are published at certain intervals to the Tivoli® Performance Viewer. You can use this
viewer to watch the results as they occur on your system, and, optionally,
log the results to a file that can be later viewed and analyzed within
the same viewer.
For Application Response Measurement (ARM) statistics,
use the administrative console Request Metrics section to specify
and the statistics you want to monitor.
- Common Base Events for problem determination and business process
monitoring
- You can specify, at the time you create an application, to monitor
service component event points — along with a certain level of detail
for those events — on a continual basis after the application is deployed
on a running server. You can also select event points to monitor after
the application has been deployed and the events invoked at least
once. In both cases, the events generated by monitoring are fired
across the Common Event Infrastructure (CEI)
bus. These events can be published to a log file, or to a configured
CEI Server database. WebSphere® ESB supports
two types of Common Base Event enablement for problem determination
and business process monitoring:
- Static
- Certain events points within an application and their level of
detail can be tagged for monitoring using WebSphere Integration Developer tooling.
The selections indicate what event points are to be continuously monitored,
and are stored in a file with a .mon extension that is distributed
and deployed along with the application. When WebSphere ESB has
been configured to use a CEI server, the monitoring function begins
firing service component events to a CEI server whenever the specified
services are invoked. As long as the application is deployed on WebSphere ESB, the
service component event points specified in the .mon file is constantly
monitored until the application is stopped. You can specify additional
events to be monitored in a running application, and increase the
detail level for event points that are already monitored. But while
that application remains active you cannot stop, or lower the detail
level of, the monitored event points specified by the .mon of the
deployed application.
- Dynamic
- If additional event points need to be monitored during the processing
of an application without shutting down the server, then you can use
dynamic monitoring. Use the administrative console to specify service
component event points for monitoring, and set detail level for the
payload that will be included in the Common Base Event. A list is
compiled of the event points that have been reached by a processed
service component after the server was started. Choose from this list
individual event points or groups of event points for monitoring,
with the service component events directed either to the logger or
to the CEI server database.
The primary purpose of the Dynamic
enablement is for creating correlated service component events that
are published to logs, which allow you to perform problem determination
on services. Service component events can be large — depending on
how much data is being requested — and can tax database resources
if you choose to send events to the CEI server. Consequently, you
should publish dynamically monitored events to the CEI server only
if you need to read the business data of the events, or if you otherwise
need to keep a database record of the events. If, however, you are
monitoring a particular session, then you need to use the CEI server
database to access the service component events related to that session.