recovery database
An issue could result when an ancillary goes down and a thread backs up the database. Setting an alert on every outbound thread to monitor the queue depth is not feasible when there are many threads.
For example, the number of messages in the recovery database nears the threshold that was set in the alert. After some time, though, the number drops below the threshold.
Because the alerts only periodically check the database, there is the possibility that the large increase in the number of threads was not recorded.
This would not be an issue since the number of threads dropped, nor is it an issue that the temporary increase was missed.
To ensure the queue does not get too large, you can now set one alert to monitor the recovery database, and not the individual threads.
To do this, use the recoverydb alert.
In this alert type, the Source can only be configured as All Threads. This is the only option on the Select Alert Source dialog box.
For example:
{ALERT
{ NAME recovery_db }
{ VALUE recoverydb }
{ SOURCE ALL_THREADS }
{ WITH -1 }
{ COMP {>= 1} }
{ FOR once }
{ WINDOW */*/*/* }
{ ACTION {
{ notify {} }
} }
}
To monitor fields, click Advanced Message dialog box.
in the Actions panel to select fields to monitor from theWhen the alert fires, the specified alert fields and their values are displayed in the Advanced Message column of the Alerts dialog box.