nsclient MEMUSE reporting different usage than server

[snapon_admin]

I'm not sure if this is an issue of the memory check counting cache in with used memory or what, but I need some assistance on this. One of our windows servers is reporting mem utilization of 71%, but Nagios says 84%. I need to know why these are different and what I need to do to make them line up better. Please see attached screens.

Task manager on server:

TM.png

Nagios:

nagios.png

[tmcdonald]

Can you post the check command being used for this service?

[snapon_admin]

Code: Select all

$USER1$/check_nt -t 60 -H $HOSTADDRESS$ -s "$ARG1$" -p 12489 -v $ARG2$ $ARG3$ $ARG4$

ARG2: MEMUSE
ARG3: -w 80 -c 90

Rest are blank.

[slansing]

Change ARG3 to this:

Code: Select all

ShowAll type=physical

And shift the warning and crit values to ARG4:

Code: Select all

-w 80% -c 90%

That should filter the memory to physical, but it will only show committed bytes. You may want to try check_nrpe:

Code: Select all

./check_nrpe –H ip.addr.of.host –c CheckMEM -a 'type=physical' '-w 80 -c 90'

We may need to take a look at your nsclient.ini file to see if you have arguments and NRPE enabled.

[snapon_admin]

k, that looks like it sort of did the trick. It's showing green now, which is good, but it still says 83% used. Is there a way to only show the used physical memory?

[slansing]

Yeah, that will have to be done through NRPE methinks, due to how that is set up with NSClient++, the other alternative would be to use NCPA where you can hand pick the metrics you want on each service through it's API.