How can I be alerted if a related group of processes runs too long?

To monitor whether a group of Processes runs too long, construct an SLC duration schedule and reference that schedule in a workflow SLC group.

About this task

The workflow SLC group contains appropriate values for server groups, servers, and other SLC match criteria. The match criteria specifies the Process, Processes, Process step, or Process steps you are interested in. For example, say that PROCESSA runs on SERVERA, PROCESSB runs on SERVERB, and PROCESSC runs on SERVERC, and that all must run within 10 minutes.

Procedure

  1. Since you are interested in knowing that multiple related Processes run within a specified amount of time, use a workflow SLC. Associate a duration SLC schedule with it.
    For this example, the duration SLC schedule has the following attributes:
    Name Schedule Type Enabled Minimum Duration Maximum Duration
    Duration Duration Schedule Yes 000:00:00 000:10:00
  2. The workflow SLC that references this duration SLC schedule is defined as follows:
    Name Enabled Generate notification if event has not occurred Monitor Window Tolerances Start Monitor Window Tolerances End
    Watch for Processes A, B, and C Yes Yes 6 6

    Monitor Window Tolerances: Start and End specify when you want an SLC to begin and end monitoring. In this instance, because a duration SLC schedule is being associated with the workflow SLC, Start Monitor Window Tolerance is immaterial. The reason is that we do not care what time the Processes begin, only that they all run within a specified duration.

    The End Monitor Window Tolerance tells IBM® Sterling Control Center Monitor how long to continue watching for the end of the workflow beyond the schedule associated with the workflow SLC. You might make the value for both the Start and End Monitor Window Tolerances as short as zero hours.

  3. Specify the duration SLC schedule previously defined. The schedule specifies a minimum duration of 0 seconds and a maximum duration of 10 minutes.
  4. On the Parameters pane, the value you select for Milestone Times Relative To, Actual Start or Scheduled Start, only matters if NSR and NER or both values are specified for the SLCs milestones.

    Concurrence Count is set to 1 because we expect only one occurrence of this workflow to be executing within the monitoring window specified.

    For SLCs with SLC calendar schedules, Concurrence Count tells IBM Sterling Control Center Monitor how many occurrences of the workflow are expected to run during the scheduled time. If “Generate notification if event has not occurred” is selected for an SLC with a calendar schedule, then SLC events are generated for each instance that does not run as expected. For SLCs with SLC duration schedules, Concurrence Count tells IBM Sterling Control Center Monitor how many occurrences of the workflow might run simultaneously.

  5. On the Milestones window, specify three milestones, one for each Process (PROCESSA, PROCESSB, and PROCESSC).
  6. Define parameters that specify PROCESSA on SERVERA for the first milestone, which we call PROCA, as illustrated in the following table:
    Key Operator Value
    Server Wild card SERVERA
    Process Name Wild Card PROCESSA
  7. On the Schedule window, specify a minimum and maximum duration of 000:00:00 to 000:10:00 (0 minutes to 10 minutes) for milestone PROCA.
  8. Additional milestones, similar to the first, must be defined for PROCESSB on SERVERB, and PROCESSC on SERVERC.

    Click the Duplicate button to copy the PROCA milestone makes it easier to create milestones PROCB and PROCC.

    The following table illustrates the milestones criteria:

    # Name Dmin Dmax
    1 PROCA 000:00:00 000:10:00
    2 PROCB 000:00:00 000:10:00
    3 PROCC 000:00:00 000:10:00