Open for Voting

Allow custom properties as values for Warning & Critical thresholds in Component Monitors

I was building out some component monitors for a client today and was displaying Windows Processor Queue Length (of The Ultimate CPU Alert for Large Environments infamy!) in a table.  I realized that queue length is irrelevant unless you know how many CPUs are in a server.  This value could be gather with a WMI query on Windows servers, but then you have to manually enter it into each component.

As described in The Ultimate CPU Alert for Large Environments we are using a custom SQL query to count the number of CPUs in a server and update a node custom property (n_cpucount, for example) with that value each day.  Sure, we use it in an alert but wouldn't it be awesome if you could reference the SQL or (preferably) SWQL version of that custom property in a SAM component and make the thresholds dynamic?!?

After thought: Adjust threshold settings and apply baseline data allows you to use the variable ${USE_BASELINE}. I wonder if a similar variable exists or if you can already cheat and use a SWQL reference?  Oh my -- I need to test this emoticons_happy.png

  • No, unfortunately not.  After pondering on it a bit, I realized that the variable ${USE_BASELINE} was built specifically for this purpose.  Perhaps if tdanner​ and team worked with aLTeReGo​ and team to allow a variable other the ${USE_BASELINE} then we could get the SWIS folks to give us a new SWIS variable that would allow us some other value. I would love to see the equivalent of the Custom Poller transform for thresholds!  (Hint, hint!)

  • Any luck with the SWQL query?

    What I'm looking at doing is pull two snmp values, i.e. throughput max and actual throughput.  The actual throughput would then be turned into a percentage by dividing actual by max.  The thresholds would then be configured at a percentage level (70% warning, 90% critical or whatever).