Administration of the Central Component › Batch Jobs › Job APCXJSMF › Alert Part

Alert Part

The alert part of APCXACAL manages all statistical exceptions and the threshold alerts. The statistical exceptions have two main parts:

• First, if this job step is marked as a runaway situation, the statistical values are saved in the alert area of the recorded job step and the job step frequency is checked. If there is a high frequency for the job step, that is, the job step is executed more than three times each week, Performance Management Assistant records this information internally as a possible alert and marks the job step as critical only. Otherwise, if this job step is not a high frequency job step, the alert is generated immediately with a corresponding pending state code.
• Second, the job step alert state is checked. If there is an alert, the measured values are checked with the runaway test and if a runaway is detected again, the alert is changed to a real one if the frequency of the job step is high. Otherwise, a new alert is created only when the runaway test against the alerted values is positive, that is, an alerted situation is imminent. If the runaway test is negative, the alert is closed.

To prevent bulk alerts if the system is overloaded, automatic alert generation is limited to five alerts in sequence for job steps with runaway elapsed times, that is, the workload of the system is in the red area. In this event, an online message appears to the first TSO user entering the online dialog.

In case of threshold alerts, a pending alert is generated immediately and the statistical values are saved in the alert area too.