Configuration & Administration

Expand all | Collapse all

Per interface utilization based threshold

  • 1.  Per interface utilization based threshold

    Posted 12-09-2019 10:59 PM

    Hi Support,
    I am trying to implement network utilization per interface basis.  Something like for filesystem as mentioned on zenoss wiki.

    Zenoss discovers virtual interface along with  physical but speed assigned to virtual interface have no relation with data processed by them due to which threshold defined based on utilization  is generating false alarm.

    Overriding template to define threshold per interface is not  feasible so i need expert help to use cproperties  (or any alternate method) and define threshold per interface basis and exclude some unwanted interface's.

    Any help/pointer/link in this regard will be much appreciated .


  • 2.  RE: Per interface utilization based threshold

    Posted 12-19-2019 02:25 AM
    Strange to see no one ever faced this issue before, still hopeful.


  • 3.  RE: Per interface utilization based threshold

    Posted 12-20-2019 01:45 AM
    Hello Alok,

    Depending on your case, you could solve this with different approaches:
    - Implement the thresholds interface per interface, but not with the Zenoss UI. Write your own scripts to automate this. Personally, I'm using Excel sheets to customise my Templates and then I use a script to implement those customisations. This has also the advantages that you have a documentation of your customisations and some sort of backup. This solution could also be enhanced with a Web UI and with a "keyword" feature to implement Template based more dynamically. Of couse, the Excel sheet is just a format and this can easily be changed.
    - Manage your thresholds dynamically within an Event Transform. The idea is that you trigger an event in all case and then filter with the help of the Transform. I guess that within the Transform you can access the cProperties, if needed. This is more probably more dynamic. However, depending on your case, it could be more resource-consuming for Zenoss.



    Laurent Hemeryck
    Monitoring Engineer