The first step in setting up your monitoring infrastructure is to determine which conditions need to be monitored and hence are the source of events. To prevent an inordinate number of extraneous events from being generated, thus reducing system and administrator overhead, you need to determine what is of interest to you and enable monitoring based on your requirements. You can leverage Enterprise Manager features such as Administrations Groups to automatically apply management settings such as monitoring settings or compliance standards when new targets are added to your monitored environment. This greatly simplifies the task of ensuring that events are raised only for those conditions in which you are interested. For more information, see Using Administration Groups.

Example: You want to ensure that the database containing your human resource information is available round the clock. One condition you are monitoring for is whether that database target is up or down. If it goes down, you want the appropriate person to be notified and have them resolve the problem as quickly as possible. Other conditions that you may want to monitor include performance threshold violations, any changes in application configuration files, or job failures. Working with events, you are monitoring and managing individual targets and issues directly related to those targets. For example, you monitor for individual database availability, individual host threshold violations such as CPU and I/O load, or perhaps the performance of a Web service.

In general, if you are primarily interested in availability and some key performance related metrics, you should use default monitoring templates and other template features to ensure the only those specific metrics are collected and events are raised only for those metrics.

Job Events: The status of a job can change throughout its lifecycle - from the time it is submitted to the time it has executed. For each of these job statuses, events can be raised to notify administrators of the status of the job.

As a general rule, events should be generated only for job status values that require administration attention. These job status values include Action Required and Problem status values such as Failed or Stopped. However, in order to avoid overloading the system with unnecessary events, job events are not enabled for any target by default. Hence, if you would like to generate events for jobs, you must: