1.1.4 Notifying users of problems in Microsoft SQL Server operation
In addition to using performance data collected by PFM - Agent for Microsoft SQL Server to display Microsoft SQL Server performance as reports, you can also use it to warn the user of a problem or error occurring during Microsoft SQL Server operation.
Suppose that you wish to notify the user by email whenever the percentage of cache hit rate is less than 85%. You can do this by setting percentage of cache hit rate is less than 85% as the abnormal condition threshold, and setting the system to send an email to the user when this threshold is reached. What the system does when the threshold is reached is called an action. The following types of actions are available:
-
Sending an email
-
Executing a command
-
Issuing an SNMP trap
-
Issuing a JP1 event
The definition of a threshold or action is called an alarm. A table of defined alarms is called an alarm table. Once an alarm table is defined, it is associated with PFM - Agent for Microsoft SQL Server. Associating an alarm table with PFM - Agent for Microsoft SQL Server is called binding. Once an alarm table has been bound to PFM - Agent for Microsoft SQL Server, whenever the performance data collected by PFM - Agent for Microsoft SQL Server reaches the threshold defined as an alarm, the event is reported to the user.
By defining alarms and actions, you can detect Microsoft SQL Server problems at an early stage and take appropriate action.
For details about how to set alarms and actions, see the chapter on alarm-based operation monitoring in the JP1/Performance Management User's Guide.