ALERTS

Purpose

The ALERTS statement specifies the conditions under which Tivoli Workload Scheduler for z/OS should generate an alert. You can specify this statement for a tracker, controller, or standby controller. You can use these alert actions when an alert condition occurs:

Send a generic alert to NetView®.
Write a message to the Tivoli Workload Scheduler for z/OS message log.
Generate a WTO (write-to-operator) message.
Send an alert to IBM® Tivoli® Monitoring (Tivoli Enterprise Portal).

ALERTS is defined in the member of the EQQPARM library as specified by the PARM parameter on the JCL EXEC statement.

Format

>>-ALERTS--+------------------------------------+--------------->
           |             .-,--------------.     |
           |             V                |     |
           '-GENALERT(-----+-DURATION---+-+---)-'
                           +-ERROROPER--+
                           +-LATEOPER---+
                           +-OPCERROR---+
                           '-QLIMEXCEED-'
 
>--+--------------------------------+--------------------------->
   |       .-OPCERROR-----------.   |
   |       | .-,--------------. |   |
   |       | V                | |   |
   '-MLOG(-+---+-DURATION---+-+-+-)-'
               +-ERROROPER--+
               +-LATEOPER---+
               +-QLIMEXCEED-+
               '-RESCONT----'
 
>--+------------------------------------+----------------------->
   |             .-,--------------.     |
   |             V                |     |
   '-MONALERT(-----+-DURATION---+-+---)-'
                   +-ERROROPER--+
                   +-LATEOPER---+
                   +-OPCERROR---+
                   +-QLIMEXCEED-+
                   +-SPECRES----+
                   '-WLMOPER----'
 
>--+--------------------+--------------------------------------->
   |          .-YES-.   |
   '-MONOPER(-+-NO--+-)-'
 
>--+-------------------------------------+---------------------->
   |             .-NETVALRT----------.   |
   '-RECEIVERID(-+-Alert Receiver ID-+-)-'
 
>--+-------------------------------+---------------------------><
   |        .-,--------------.     |
   |        V                |     |
   '-WTO(-----+-DURATION---+-+---)-'
              +-ERROROPER--+
              +-LATEOPER---+
              +-OPCERROR---+
              +-QLIMEXCEED-+
              '-RESCONT----'

Parameters

GENALERT(error condition,...,error condition): Defines the conditions under which Tivoli Workload Scheduler for z/OS should send a generic alert to NetView.
MLOG(error condition,...,error condition|OPCERROR): Defines the conditions under which Tivoli Workload Scheduler for z/OS should write a message to the message log.
MONALERT(error condition,...,error condition): Defines the conditions under which Tivoli Workload Scheduler for z/OS sends a generic alert to the IBM Tivoli Monitoring agent. This parameter is significant only if the MONOPTS statement is provided.
MONOPER(YES|NO): Defines if the following error conditions specified in the MONALERT parameter are in effect for monitored jobs only (default) or for all jobs. The error conditions are ERROROPER, LATEOPER, DURATION, and WLMOPER.
RECEIVERID(Alert Receiver ID|NETVALRT): Defines the NetView alert receiver that generic alerts are sent to. Specify this keyword if the alert receiver in NetView address space that handles Tivoli Workload Scheduler for z/OS alert automation does not have the NetView default ID, NETVALRT.
WTO(error condition,...,error condition): Defines the conditions under which Tivoli Workload Scheduler for z/OS should generate a write-to-operator (WTO) message.

Error conditions

You can specify one or more of the following error conditions for each alert type. Note that only OPCERROR and QLIMEXCEED are applicable to a tracker; other error conditions are ignored if you specify them.

DURATION

The alert action is taken when an operation in the current plan is active for an unexpectedly long time. This means that an operation that has started (extended status S) must be active longer than its estimated duration multiplied by either the following values:

The alert action limit that you set for ALEACTION (for details, see JTOPTS).
-OR-
The duration feedback limit that you set for LIMFDBK (for details, see JTOPTS). This value is used if ALEACTION is not set.

then divided by 100.

For example, if an operation has an estimated duration of 10 minutes and the limit for the alert action is 200, the alert action is taken if the operation is active for longer than 20 minutes. The alert action is also taken if the operation has been started but the associated job or started task has not yet started to run after 10 minutes (no A2/B2 event has been received), that is, the operation has had status/extended status SU and SQ totally for more than 10 minutes. The alert action is taken only for operations that have started status.

For MLOG and WTO alert actions, message EQQE028I is issued for an operation at a general workstation and EQQE038I for an operation at a computer or printer workstation for long running operations. Message EQQE039I is issued for computer operations that have been submitted but have not started.

Notes:

The value used to select the operations for which a long duration alert must be issued is set with the ALEACTION keyword in the JTOPTS statement. If ALEACTION is not specified, the value set for LIMFDBK is used instead. In this case, the value for the feedback limit that you can optionally enter in the application description is ignored.
If the alert action limit is 0 or the alert action limit is not specified and the feedback limit is 0, the alert action is taken as soon as the operation is active longer than its estimated duration.
If the estimated duration of an operation is 99 hours, 59 minutes and 01 seconds, no duration alert is sent for this operation.

ERROROPER

The alert action is taken when an operation in the current plan is set to ended-in-error status. For MLOG and WTO alert actions, message EQQE026I is issued for an operation at a general workstation, and EQQE036I for an operation at a computer or printer workstation.

LATEOPER

The alert action is taken when an operation in the current plan becomes late. An operation is considered late if it reaches its latest start time and does not have the status started, complete, or deleted. For MLOG and WTO alert actions, message EQQE027I is issued for an operation at a general workstation and EQQE037I for an operation at a computer or printer workstation.

Note:

Use LATEOPER only when deadlines are accurate because it can affect the performance of Tivoli Workload Scheduler for z/OS.

OPCERROR

The alert action is taken when a Tivoli Workload Scheduler for z/OS subtask or Tivoli Workload Scheduler for z/OS subsystem ends unexpectedly. For MLOG and WTO alert actions, message EQQZ045W and EQN019E are issued. If GENALERT action is specified and the EQQN019E alert condition occurs, then the subtask-failed alert is sent to NetView.

Note:

OPCERROR is always in effect for MLOG.

QLIMEXCEED

The alert action is taken each time a Tivoli Workload Scheduler for z/OS subtask queue exceeds a threshold value. Except for the event-writer queue, the threshold values are multiples of 10 between 10% and 90%, and then 95% and 99%. Tivoli Workload Scheduler for z/OS checks the size of a queue when an event is added to it. Except for the event-writer queue, Tivoli Workload Scheduler for z/OS subtask queues can contain up to 32 000 elements.

The size of the event-writer queue is determined by the ECSA you allocate. The queue is checked each time the event writer is about to read events; the alert action is taken if the queue exceeds 50%. If the event-writer queue becomes full, a message is issued indicating how many events have been lost.

The value in the alert shows the actual percentage used, which will be more than the threshold value. For MLOG and WTO alert actions, message EQQZ106W is issued.

RESCONT

You can specify RESCONT (resource contention) only for MLOG and WTO alert types. The alert action is taken when an operation has been waiting on a resource queue for the time specified on the CONTENTIONTIME keyword of RESOPTS. Message EQQQ515W is issued.

SPECRES

The alert action is taken when the time that an operation in the current plan is waiting to allocate a given resource exceeds the time specified by the RESOPTS CONTENTIONTIME parameter. This alert takes effect when it is defined in the MONALERT parameter.

WLMOPER

The alert action is taken when an operation in the current plan is promoted by WLM. The alert is sent only if specified in the MONALERT parameter.

Examples

 ALERTS MLOG(ERROROPER,LATEOPER,DURATION)   1 
        WTO(DURATION)                       2 
        GENALERT(ERROROPER)                 3 
        MONALERT(DURATION,OPCERROR,WLMOPER) 4 
        MONOPER(YES)                        5

In this example of an ALERTS statement:

1: Tivoli Workload Scheduler for z/OS writes a message to the message log for operations that are set to ended-in-error status, are late, or are active an unexpectedly long time. Although it is not specified, the OPCERROR condition also applies for the MLOG alert action.
2: A write-to-operator message is generated for operations that are active an unexpectedly long time.
3: Tivoli Workload Scheduler for z/OS sends a generic alert to NetView for operations that are set to ended-in-error status. By default, generic alerts are sent to the alert receiver NETVALRT.
4: Tivoli Workload Scheduler for z/OS sends a generic alert to IBM Tivoli Monitoring for operations that have long durations, for Tivoli Workload Scheduler for z/OS substaks or subsystems that end unexpectedly, and for operations promoted by WLM.
5: Tivoli Workload Scheduler for z/OS sends a generic alert to IBM Tivoli Monitoring only for monitored operations (operations with EXTERNAL MONITOR = YES) satisfying the conditions specified in the MONALERT parameter and not for all jobs.