Open for Voting
over 3 years ago

If...Then...Else statements for Alert Manager

I think it would make managing alerts much easier if we were able to build an if-then-else statement into the actions for an alert.

My example...

I create an alert for a particular set of criteria, but I want to enable a different response depending on who is responsible for solving the issue.

IF Node.Owner = "Server" THEN do(Action1)

IF Node.Owner = "Network" THEN do(Action2)

I gather it would be fairly complicated to implement, but I think it would really improve the alert manager.

Parents
  • I owe you folks some news here.  This has been our #1 feature request for some time.  We've done significant research on this FR including surveys and 1:1 interviews with many of the people who voted for and against it.  It turns out that while there are 1400+ votes for this feature, they're votes for several different distinct things.  The two most popular things are actually contradictory: make individual alerts simpler to configure vs make each alert more powerful by making it more modular and complex.

    We're still researching this but I have not yet found a solution that does justice to at least most of the people voting for this FR.  You can help though.  If/when you vote for this or if you already have, add a comment with your thoughts on what problem you're trying to solve by voting for this feature and a rough thought of what a good solution would look like to you.  As the community discusses why we agree or disagree, we will get closer to understand the root problem(s) and the solution(s) will start becoming more clear as well.  Or wait longer and I'll figure it out myself!  emoticons_wink.png

Comment
  • I owe you folks some news here.  This has been our #1 feature request for some time.  We've done significant research on this FR including surveys and 1:1 interviews with many of the people who voted for and against it.  It turns out that while there are 1400+ votes for this feature, they're votes for several different distinct things.  The two most popular things are actually contradictory: make individual alerts simpler to configure vs make each alert more powerful by making it more modular and complex.

    We're still researching this but I have not yet found a solution that does justice to at least most of the people voting for this FR.  You can help though.  If/when you vote for this or if you already have, add a comment with your thoughts on what problem you're trying to solve by voting for this feature and a rough thought of what a good solution would look like to you.  As the community discusses why we agree or disagree, we will get closer to understand the root problem(s) and the solution(s) will start becoming more clear as well.  Or wait longer and I'll figure it out myself!  emoticons_wink.png

Children
  • cobrien​ with respect to complexity; I see no reason why both are not desirable. On the one hand you have better out of the box alerts; maybe better examples (canned alerts are nice but with limitations) . On the other hand you have better granularity with scheduling of alerts; much like what was in alert central and webhelpdesk. It's simply catering to both crowds; helping the novice expand their knowledge and helping the advanced user expand their functionality.

  • cobrien​, you  have the simpler  approach now...  Maybe a newbie wizard for the simple stuff.

    The feature request is to allow those needing a more advanced bit of logic to determine if and when an alert/notification needs to go out has options to build much more complex constructs. 

    Many products out there have 2 or more ways to edit monitors and or alerts...use the wizard or use the raw (got to know what you are doing interface).

  • Please see my post of 8 March 2017. I think this summarises my main issues.
    I have no problems using custom properties within an action to affect the behaviour of that action. E.G. use CP to derive email addresses for email alerts.
    What I can't do is use a CP (or other field) to determine which action(s) to perform.
    If I had this capability, I would only need for example one node down alert for all my needs. The conditional actions would give me all of the other functionality I need.




  • Sounds like there may be room for a 'option' to select optional or more advanced querying structure... But then again too many 'options' via radio buttons and drop downs can take you slightly outside of the user friendly element that SW seems to be trying to keep here.