3.3.5 Considerations for monitoring the alarm status in units of instances
- Organization of this subsection
(1) Allowing alarm events to show the instance that caused an alarm
When a monitoring agent collects performance data in units of instances, define alarms so that information that indicates the instance that caused an alarm is shown in the message text of an alarm event.
The following is an example of how to define an alarm.
Suppose you want to use PFM - Agent for Platform (Windows) to monitor disks. You want PFM - Agent for Platform to send a warning alarm when the remaining free disk space becomes less than 30% and an abnormal alarm when the remaining free disk space becomes less than 10%. Define the alarm as follows so that the message text in the alarm event indicates which instance caused an alarm.
-
In the window for specifying the main Information for the alarm, specify the alarm message text as follows:
Alarm message: Disk %CVS1#:Available disk space is %CVS2%
#: %CVS1 is replaced by a measurement value in the field specified for the first condition in the alarm conditions in step 2.
-
Specify the alarm conditions.
-
First condition
Specify the field that indicates the name of an instance.
Record: PI_LOGD
Field: ID#
Condition: =
Abnormal value: *
Warning value: *
#: Key field of the PI_LOGD record (information for identifying an instance)
-
Second condition
Specify the warning and abnormal threshold values for the remaining disk space indicated by %CVS2.
Record: PI_LOGD
Field: % Free Space
Condition: <
Abnormal value: 10
Warning value: 30
-
As a result, in the message text, variable %CVS1 is replaced by the volume name of a logical disk and variable %CVS2 is replaced by a measurement value in the % Free Space field.
Example message text: Disk C::Available disk space is 29.0123%
(2) Allowing JP1 events to show the instance that caused an alarm
When Performance Management is linked with the integrated management product (JP1/IM) to monitor the operating status and a monitoring agent collects performance data in units of instances, define alarms as shown below. This will cause the instance that caused an alarm to be displayed in the message area for a JP1 event.
The following procedure is an example of how to define an alarm and its corresponding JP1 event so that the JP1 event collected by JP1/IM indicates the instance that caused the alarm.
-
Define the alarm.
For the procedure, see 3.3.5(1) Allowing alarm events to show the instance that caused an alarm.
-
In the JP1 event settings, specify the message shown below. This will cause the alarm message text to appear in the message area in the JP1 event settings.
Message: %MTS
(3) Checking the status of an instance from an alarm event
You can associate an alarm definition with a report. While doing so, if you associate a report specified to display the same information as for the fields that are referenced for alarm conditions, you can check the status of an instance with ease.
We take the example described in 3.3.5(1) Allowing alarm events to show the instance that caused an alarm. In this example, a report that displays the values of % Free Space and ID fields is associated with the alarm for which the % Free Space and ID fields in the PI_LOGD record are referenced for alarm conditions.