Administration of the IMS Feature › Overview › Alert Management

Alert Management

In IMS there are hundreds of transactions but not all of these transactions need performance monitoring. Many transactions, even though they might be high consumers, only execute periodically and do not warrant concern.

Each time the IMS Feature executes, important top consuming transactions can be identified and alerted automatically. This process works in two steps:

1. The top number of consuming transactions are identified based on the scope you define as a parameter, for example, TOP Limit = 10.
2. Within this TOP Limit, the current execution consumption values of the transactions are compared to the statistical information that is maintained for the same transactions on the database. If the actual consumption exceeds the statistical limits, an alert is issued automatically. This alert is referred to as a statistical alert.

   Additionally, the user can manually issue alerts explicitly for transactions that use the online Alert Management option. These alerts are referred to as user alerts.

The alert management option provides all the information that is necessary to manage the alert. Information is provided in the form of state and reason codes that identify the situation. For details about how to review alerts online, see the User Guide that corresponds with your measurement tool.

All alerted transactions are also listed in the APCSALT report file of job APCYJLNA. (For details regarding this job, see Jobs APCYJLST and APCYJLNA - Process CA MAT Measurement). Listed for each alert transaction is the average of CPU time, the current CPU time, the number of relevant history entries of the current transaction, and the resulting standard deviation in CPU time. Depending on the parameter setting for online alerting, these CPU times either are in percent as in previous releases or are the transactions' absolute CPU time. Using the APCSALT report file, user written REXX procedures can use the key word parameters to process the alert information, for example, pass information to change management or provide programmer notification of alerts.

The example following shows the APCSALT report file output for a runaway based on CPU time percent.

$$ALERT                                                                
$SN=IMS123                             SYSTEM NAME                      
$TX=OZN31000                           TRANSACTION NAME                 
$CM=00339                              MEASURED CPU TIME %              
$CA=00052                              AVERAGE CPU TIME %               
$AT=005                                ALERT TEXT WITH NNN LINES        
2010-01-09 PMA ALERT ID 16193    BY STAT       
-------------------------------------------------------------------------
           TRANSACTION : OZN31000   SYSTEM : IMS123     PGM : IMSPR001  
           CPU%  MEAS  :   3.39   AVERAGE :    .52   STD.DEV :    .72   
-------------------------------------------------------------------------
$$ALEND
$$ALERT                                                                 
$SS=IMS                                SUBSYSTEM ( CICS / IMS )         
$SN=IMS123                             SYSTEM NAME                      
$TX=OZN31000                           TRANSACTION NAME                 
$CM=00339                              MEASURED CPU TIME %              
$CA=00052                              AVERAGE CPU TIME %               
$AT=005                                ALERT TEXT WITH NNN LINES        
2010-01-09 PMA ALERT ID 16193    BY STAT       
-------------------------------------------------------------------------
           TRANSACTION : OZN31000   SYSTEM : IMS123     PGM : IMSPR001  
           CPU%  MEAS  :   3.39   AVERAGE :    .52   STD.DEV :    .72   
-------------------------------------------------------------------------
$$ALEND