Ad Widget

Collapse

Alert based on age of event?

Collapse
X
 
  • Time
  • Show
Clear All
new posts
  • kehall
    Member
    • Sep 2008
    • 30

    #1

    Alert based on age of event?

    Is it possible to have a condition in the main criteria "event.age" or something similar - that way you can send alerts based on triggers that have been active for e.g. 90secs or greater? This could cut down on alerts where a problem is very short lived (e.g. too many processes for 30 seconds).

    Keith
  • Alexei
    Founder, CEO
    Zabbix Certified Trainer
    Zabbix Certified SpecialistZabbix Certified Professional
    • Sep 2004
    • 5654

    #2
    You may use escalations for this, just do not define any actions for the first step or first N steps.
    Alexei Vladishev
    Creator of Zabbix, Product manager
    New York | Tokyo | Riga
    My Twitter

    Comment

    • kehall
      Member
      • Sep 2008
      • 30

      #3
      Thanks... Unfortunately this doesn't work as expected...

      I created a new alert:
      Enabled escalations
      Set period to 300 secs
      Enabled recovery message
      Set Condition Trigger severity = Average

      Action step from 2, to 2, period 0... send message to single user, no conditions.

      No other actions..

      I turned off a service to produce an Average event, then turned it back on after about 120 seconds.

      No email during that time... so that bit is ok.

      However, 300 seconds after the original event I notice Actions in Events showing "in progress" and I get an email (not right as the fault status is no longer true)!

      Now Date/Time : 2008.10.16 - 10:30:28
      Event Date/Time: 2008.10.16 - 10:25:26 (age 5m )
      Trigger : Windows Time Service not running
      Status : ON
      Value : 0
      Severity: Average


      Not only that, but 300+120 seconds later I get another email (but it wasn't the recovery message as that has a different subject line)

      Now Date/Time : 2008.10.16 - 10:32:26
      Event Date/Time: 2008.10.16 - 10:27:26 (age 5m )
      Trigger : Windows Time Service not running
      Status : OFF
      Value : 0
      Severity: Average


      So it's not doing what is required.

      If it were possible to set a condition in the action "where status = off" then that would stop the first email I think, but then I think the 2nd email would be sent... I don't think the recovery message should be sent unless a problem message had been sent out.

      To be honest, IMHO the recovery message should be immediate not delayed, so not sure why it's delaying an 'OK' status email too?

      Thoughts?

      Regards,

      Keith.

      Comment

      • kehall
        Member
        • Sep 2008
        • 30

        #4
        Just wondering if there is a workaround or whether it is a bug, feature enhancement or what, please?

        At the moment without more control on the condition of actions and escalations etc, emails are flying in all the time and thats a problem because you become immune to "proper" issues then!

        Keith

        Comment

        Working...