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.

How to setup an alert for a single switch CPU

Hello,

Our Cisco catalyst switches have multiple CPU's.  NPM is reporting the average CPU utilisation across all switches in the stack.  I'd like to setup an alert to trigger if any single CPU is equal to or above 85%.  I've used a Custom SQL Alert (Advanced), selected 'Node' and used the following script;

join CPUMultiLoad on Nodes.NodeID = CPUMultiLoad.NodeID

where DATEDIFF(mi, CPUMultiLoad.TimeStampUTC, SYSUTCDATETIME()) < 5

and CPUMultiLoad.AvgLoad > 85

However this only seems to pickup the CPU on our servers, it doesn't trigger for switches.

Can anyone help please?

Thank you!

  • I don't have a stack switch handy to test on, but if you are finding that this alert works for servers and not switches it could be that stacks are storing their loads on a different table.  Can you search the cpumultiload view for the nodeid of one of your stacks to confirm it shows up in there?  Also, it might be a more efficient search to query cpumultiload_current (double check that table name, as i don't have the db in front of me) if you are only interested in the most recent polling intervals.

    If you don't find the switch stacks on that table then you would want to dig around and see if there is maybe a stack specific table or one with a list of stackmembers.

  • How do I search in the cpumultiload view? 

    Thanks

  • Ah I assumed you had made that query yourself.   There is a tool on your orion server called database manager that you can use to browse the tables and look for things,  for this purpose I think the query would be select * from cpumultiload where nodeid= (your stack nodeid)

  • Thanks for the directions, I can see the following which only shows NodeID 47;

    pastedImage_0.png

  • And 47 is a stack switch I'm assuming? It looks like only a single cpu index is showing up there,  which would indicate to me that the separated values are being written to another DB table away from "normal" nodes, now the trick would be to find which table that is.   Next time I'm working with a client that has stack switches I'll try to find it,  unless someone else can chime in before then.

  • Can anyone advise on how to alert if a single CPU utilisation goes above 85% please?  At present NPM alerts only on the "average" CPU utilisation across a stack.  Hopefully this should be straight forward as the data is shown in the Switch Stack section;

    pastedImage_0.png

    Thank you.