This discussion has been locked. The information referenced herein may be inaccurate due to age, software updates, or external references.
You can no longer post new replies to this discussion. If you have a similar question you can start a new discussion in this forum.

Advanced Alerts and site bouncing notifications?

Is there a way to define an advanced alert to send an email notification if WAN sites are bouncing?  i.e.  I want to send an email notification if any specific site goes DOWN/UP more than ...5 times in 60 seconds.  Rather than get 5 separate emails stating DOWN/UP, I'd like to receive a secondary notification that informs my staff a site is suffering a chronic issue.  I don't see any way to set a threshold of time on an advanced alert.
  • Chris... First, you can use the timers in the advanced alerting engine to delay the rearm for a period of time, which in effect correlates or groups the down events for that period of time. Naturally, this might give the appearance that a device is hard down for a longer period of time than it really is, but it serves the purpose of reducing the number of notifications. Now for the question about chronic issues, like your other question about "anytime changes", the only way I know to do it will require some custom SQL configuration in order to make it happen. You can create a job on the SQL server that counts the number of "Node Down" events for each device over a given period of time and have it populate a custom property with these values. Now, you can create an alert that monitors this custom property and triggers when one of these values goes above a predetermined number. Again, not the ideal situation, but it will do until something better comes along. This can be done not only for node down events, but any other event as well, but keep such thinigs in mind as how long you're storing your events, as this will have an impact on the time periods available for use with this scheme. Hope this sparks some ideas for you... ~ Vic
  • I know this is an insanely old post but i'm not sure if this was ever solved or available as a new feature.  I am still trying to alert for bouncing circuits, is there a way to somehow create a rule for

    if node down = >3, in minutes = < 20 then send alert. 

  • Possible but not easily.

    Only two ways I can think off currently are:

    1) One alert for down that has a trigger action of a custom VB/PERL/etc. script that increments a property of the node TimeDown20, and a reset action that has a similar script decrements after it becomes stable for 20min. Then another alert that triggers whenever the TimeDown20 >= 3.

    This will require you to be comfortable with scripts and have permission to change things in your database through them.

    2) One alert that triggers for down instantly (or rather within 1 min due to how SW detects down nodes).

    Use a custom SQL trigger that looks into the alert messages tables for the last 20 minutes and see if the count() of the down alert >= 3.

    Somewhat easier in practice, but I suspect this would be much more resource intensive as you need to search through all alert messages (as you can't have multiple active alerts for the same alert).

    Kind of busy so someone else will need to help you set up one or the other if you choose to implement one of them.