NSClient++ Help (#1) - Averaging Counter values? (#225) - Message List

Averaging Counter values?

Hello,

I am looking to monitor network total bytes (via Counters), but I want to try and filter out spikes so I am not alerted if I happen to take a sample when network activity has spiked. I want to be alerted when it is over a value for a certain period of time. So, take some type of average.

Is this somehow possible with NSclient++? Would this be use to other people or am I going about this all wrong.

I would think that maybe Nagios could somehow do the averaging (or not altering until xx passive checks come back as failed), but I couldn't find out how to accomplish this.

Any ideas?

Thanks,

-jr

  • Message #652

    there are time filters for CPU load for instance, but I have no idea if it works on other WMI counters that way.

    check_nrpe -H $HOSTADDRESS$ -u -p 5666 -t 60 -c CheckCPU -a warn=$ARG1$ crit=$ARG2$ time=$ARG3$ time=$ARG4$
     time=1m
    

    In case you figure something out let me know

    • Message #655

      Right. I have noticed that. Didn't see anything similar for other generic counters/WMI values. I think it would need to capture data on a constant basis and then compute average when it came time to send the data.

      I guess its not that important after all. If I can ever get it built in visual studio I'll see if I can dive in and get it done.

      -jr

      • Message #657

        Id say it is hard to do out of the box, I have had ideas along this lines (ie. allow you to add "your own" cpu checks on other counters) but noone has ever requested it so it has not happend... request it and I shall try to add it... but probably not "soon"...

        MickeM

Subscriptions