11.3.3 Designing error detection and reporting
After considering the monitoring methods and targets, consider what action to take when a problem is detected.
|
- Organization of this subsection
(1) JP1 events and event levels
JP1/IM uses JP1 events to monitor events occurring in the system. Each JP1 event is assigned an event level that indicates its severity.
Consider how to detect and deal with problems in the system according to their event level.
- Categorizing JP1 events by event level:
-
Categorize JP1 events based on their event level (attribute name: E.SEVERITY).
JP1 events define an event level associated with a specific event that occurs in the system. Each product that issues JP1 events sets an appropriate event level for the types of events it manages.
When a SNMP trap or entry in the Windows event log is converted to a JP1 event, its severity is associated with a JP1 event level. If you use the integrated monitoring database and change the event level of a JP1 event to one defined by the user, the user-defined event level is associated with a JP1 event level. To monitor these JP1 events, categorize them based on the severity of the source event. When a message in a log file is converted to a JP1 event, the event level of the resulting JP1 event is determined by the converter settings. Set the converters to assign an appropriate event level.
- Checking categorized JP1 events:
-
Check whether the JP1 events are in the appropriate category for their event level.
The event levels assigned by products in the JP1 series are generally suitable for normal operation. However, check whether reclassifying some events under another event level would suit your system operation better. For example, you might want to exclude some events that are normally classed as severe events.
(2) Response based on urgency
For the categorized JP1 events, consider response procedures based on the urgency of the event.
The following describes an example of response procedures for events with three levels of urgency (Urgent, Severe, and Normal).
- Events that require immediate attention:
-
Consider whether any of the events managed by JP1/IM require an urgent response. A system failure, for example, requires an immediate response because of the far-reaching consequences to work tasks.
Consider reporting such an error to the system administrator.
To avoid placing an undue burden on the system administrator, make sure that only JP1 events requiring a rapid response are reported. Also consider changing the range of events reported to the system administrator according to their content.
In JP1/IM, you can send an emergency notice by executing a command from an automated action. Consider how to identify events that need to be reported to the system administrator, and the command to use for notification (for example a mail sending command), and then define an appropriate automated action.
- Severe events that require monitoring:
-
Consider whether events whose event level indicates a problem should be handled as severe events that require monitoring.
Such events can be forwarded to a JP1/IM manager host where they can be centrally managed.
Use the event level or other attribute to define which JP1 events are to be handled as severe events in JP1/IM. On the Severe Events page of the Event Console window, you can view a list of the events defined as severe events and manage their response status.
- Normal events that must be checked periodically:
-
Events issued in the normal course of operation, such as events indicating that a job has ended normally, can be used in the following ways depending on your system operating requirements:
-
Keep for use as an operating history.
Example: JP1 events related to job execution can be saved as a job execution log.
-
Use to compile statistical information.
Example: The start and end times reported in JP1 events related to job execution can be used to compile statistical information on job execution times.
Utilize urgent and severe events, by displaying them in JP1/IM - View or outputting them as a CSV snapshot, when you review the system setup, for example.
You can also utilize events that are not monitored by JP1/IM by using the following commands to output the database contents on the hosts in the system in CSV format.
-
The JP1/Base jevexport command
Outputs the contents of the event databases in CSV format.
-
The JP1/IM - Manager jcoevtreport command
Outputs the contents of the integrated monitoring database in CSV format.
-
(3) Monitoring from the Central Scope
You can customize the Central Scope to display events with specific event levels.
Because a monitoring tree shows the layout of the entire system in tree form, you can easily see how the objects being monitored relate to one other. However, depending on the structure of the monitoring tree, a single failure event might cause a large number of monitoring objects to change to a color that indicates an error.
Under the default settings, changes in the status of monitoring objects are triggered by the event levels of JP1 events. However, you can customize the way in which status changes are triggered to suit your system operation.
You can also exclude monitoring objects from monitoring in the Central Scope.