Forum Discussion

drazzopardi's avatar
6 years ago

Alert suppression when collectors are unreachable


I am having a frustrating experience when I have tried to make LM redundant using multiple collectors, but then I am at the mercy of the customers internet connection so when the collectors go down, I then get a couple of thousand host status events rather than a number of critical collector alerts making it very hard to see the wood for the trees in the UI alerts page as I cannot just filter out host status alerts as I could have more from other customers.

Am I missing something?

Do others have a decent scalable workaround?

Has this been raised as a feature request before? and if so is there a reason its not materialised?

Thanks in advance


2 Replies

Replies have been turned off for this discussion
  • @drazzopardi

    I'm sorry to hear that you are experiencing issues with alerts. To start, are you currently monitoring the device the collector is installed on? If so, you can filter "Host Status" on the alerts page and the collector down alerts will populate. Alternatively, you can create a dynamic group for your collectors by leveraging "isCollectorDevice()" and filtering for that on the alerts page. 

    In terms of creating dependencies, there are currently two datasources that are mentioned on this page (below) that may be able to help. 



    Let me know if you have any further questions! 


  • Hi @Michael Fisher

    It is more the case when the collector becomes unavailable to report back to LogicMonitor. So if the collector cannot report if something is up or down, it would be clearer to alert only on that the collector or collector group was disconnected rather than showing every device on the network going down as it isn't actually accurate in the majority of cases.  I would have thought this would be achievable as you would have the collector alert first and from what I understand the host status alert 6 minutes later.

    I appreciate the dependencies or Parent/Child suggestion, however, I see that has its own strengths for a working branch office monitored scenario where the gateway goes down and we suppress everything sitting behind it.