Forum Discussion

Eugene_C's avatar
8 years ago

Run DataSource at specified time or only during specified time window

Hello,

It'd be really great to be able to run datasources by scheduler (Collect every). For example, at 3 am every day or 2 pm every Monday. 

I'm aware of pollinginterval but even if i set it to run every 3600 min (every day), i can't set time when it should be executed. 

9 Replies

Replies have been turned off for this discussion
  • Mike_Suding's avatar
    Mike_Suding
    Former Employee

    hmmm....please tell us more. What datasource/script do you need to run at this type of schedule?  thanks for the post

  • Agreed - but I would take it a step further and bake it into the SDT functionality.  "Only monitor this datasource(s) from this time to this time."  We have a number of dev environments that are shut down over the weekend and at night.  No need to waste the resources trying to get monitoring data for something we know will be down half the time.

  • Sure.

    We're providing retail management system for our customers. We're trying to replace our in-house monitoring system with LM. We have a lot of monitors that check different parameters of how retail system works (like missing receipts, etc.) We need to run this test during AM hours so that Support team could fix all errors. 

  • I also have scenarios where I would like to schedule a datasource to only run at a certain time of the day - maybe like Windows Task Scheduler. 

    Example - Pulling disk defgrag reports overnight.  With large volumes, this may take a while to execute.  I'd really only want to execute it overnight - or maybe even once a week.

    I know I can set the ".pollinginterval", but I can't make it execute at a certain time. It's still a bit arbitrary on when it does the 1st check, which then determines the time for all future checks.

  • I need this too.  As an example use case, we have CSV files that are transferred from host to host.  There is no point continually checking for the existence of the CSV file when it only needs to exist between 3am and 4am.

    Another use case, at instance level, we have certain volumes that are used as temporary space during the creation of backups (tarballs, zips, etc.).  We're only interested in monitoring these volumes during the backup period, say 11pm to 2am.  Outisde of these hours we don't care how these volumes are used.

  • have you considered doing a script based event source? You could add a date/time execution to the script that says if ! 11pm -> 2am exit, otherwise run the check, then give an event alert if the files or processes don't exist? 

    same could be true for a datasource, a script based datasource could have the same, if between a certain time just return 0 and exit, therefore not performing any actual task unless it's between your time.

    a crude workaround but works for what you're looking for at the moment. 

  • 21 minutes ago, Tom Lasswell said:

    have you considered doing a script based event source? You could add a date/time execution to the script that says if ! 11pm -> 2am exit, otherwise run the check, then give an event alert if the files or processes don't exist? 

    Tom, 

    thanks for the advice. Actually, right now i'm using this workaround. But it has some downsides.

    And LM support recommended to request this feature on the forum :)/emoticons/smile@2x.png 2x" title=":)" width="20">

  • 5 hours ago, Tom Lasswell said:

    have you considered doing a script based event source? You could add a date/time execution to the script that says if ! 11pm -> 2am exit, otherwise run the check, then give an event alert if the files or processes don't exist? 

    Hi Tom,

    The problem with this method is that our operators are not coders, so they need a UI to do the configuration work.  Also, editing scripts has higher risks than making changes via a UI.  Code would need to undergo peer review and testing, whereas with a UI you trust that the UI implementation in the product is tested and trusted.

  • 2 hours ago, Mosh said:

    Hi Tom,

    The problem with this method is that our operators are not coders, so they need a UI to do the configuration work.  Also, editing scripts has higher risks than making changes via a UI.  Code would need to undergo peer review and testing, whereas with a UI you trust that the UI implementation in the product is tested and trusted.

     

    I agree, just saying as a temporary workaround for now.