Automic Workload Automation

Expand all | Collapse all

Recommendation for making sure processes (schedules, etc) are active?

  • 1.  Recommendation for making sure processes (schedules, etc) are active?

    Posted 07-30-2015 10:19 AM
    We just had an incident yesterday where one of our Operators inadvertently canceled the schedule for one of our clients.  Give the number of clients (and therefore schedules) we automate in UC4, we didn't realize this schedule had been killed until several hours later.  Not our finest hour...

    What tactics have other folks employed to make sure this kind of thing doesn't happen?
    I know one approach would be to update our Operators' UserGroup and remove the ability to cancel schedules (or particular objects), but that may be a bit more draconian that necessary.
    Another approach could be to build a var listing various critical processes we want to ensure are running, and periodically loop through that var and throw an alert for anything that isn't running....but that would require a lot of ongoing maintenance to keep the var up-to-date, so I'm not sure if that's the best solution either.

    Any other suggestions from the masses?

    Thanks in advance!

      -- Daryl

  • 2.  Recommendation for making sure processes (schedules, etc) are active?

    Posted 07-30-2015 11:37 AM
    Depending on our requirements we have used one or more of the following methods to ensure that Schedule objects are active as required.  We not only check for Active but also ensure that it is not in a Stopped status (1563).

    1.  Time Events that execute continuously.  They, as you suggest, read a Variable with the Schedule object names that it is to monitor.  The Validity keyword is the object name and the Value is the Notification addressing for exceptions.  All checks are made twice with a couple of second delay to account for a Turnaround to handle possible a false positive.

    2. The Process tab of the Schedule activates a Time Event that monitors the Schedule that activated it.  The Event also has logic to know when it should end due to Turnaround that was passed at activation from the JSCH_PERIOD and JSCH_PERIOD_TA attributes.

    3. Cockpit Monitor that executes continuously.  It displays the counts of various schedules that are to be active and are used by our operations staff as part of their monitoring duties.

    4. Properties/Result on Schedule tasks to execute a Notification should it fail to properly activate.  We have a few Schedules that are not continuously active and are themselves tasks on a Schedule.

  • 3.  Recommendation for making sure processes (schedules, etc) are active?

    Posted 08-03-2015 01:18 PM
    I would vote for the var list that is looped through periodically.  We do something similar every 30 minutes.

  • 4.  Recommendation for making sure processes (schedules, etc) are active?

    Posted 08-03-2015 03:14 PM
    Hi Daryl,

    Yes, I also use a var list and prep_process it. But I also use a process on the DB that keeps "learning" what is being activated to complement the vara. It periodically queries the EH (Activities) table in search of events, schedules and c_periods and checks with what I have on my vara-list. If there's something on the query that is not on the vara, I send a CALL object to the administrator to query if it should be included on my vara-list (YES/NO). It's not a 100% 'always updated vara' proof but it helps me to be on top of it.