Page 1 of 1

Memory False Alert due to cached by os

Posted: Wed Jan 30, 2019 1:04 am
by axvers
Hi Support,

We received a lot of "memory usage is critical" alert these days.

And I ssh to the hosts, it seems like that os would temporarily cache available memory as it can.

I want to know what we should do to make Nagios XI skip cache memory?

Could you provide some steps? (maybe like, what plugin we need to change/ how to change the plugin instead of default one.)

target servers:
- os: all CentOS 7, 64-bit

Nagios XI Server:
- os: CentOS 7, 64-bit
- wizard: snmp linux only
- clean Nagios XI environment

Re: Memory False Alert due to cached by os

Posted: Wed Jan 30, 2019 12:04 pm
by scottwilkerson
If you are setting them up with the Linux Wizard, you can go to the CCM and edit the memory services so the $ARG2$ looks like this adding the -n flag

Code: Select all

-a '-n -w 20 -c 10'

Re: Memory False Alert due to cached by os

Posted: Thu Jan 31, 2019 9:08 pm
by axvers
Hi Support,

I added '$ARG2$' in '>_commands' in ccm;

and added the '$ARG2$' in the memory check command.

But after clicked the 'Run Check Commnand', it still show physical memory is "Critical".
(Photo1 and 2)

Re: Memory False Alert due to cached by os

Posted: Fri Feb 01, 2019 7:59 am
by scottwilkerson
My recommendation was assuming you were using the standard Linux Wizard with an agent.

Unfortunately there is no way to remove the cached memory on the SNMP wizard.

Re: Memory False Alert due to cached by os

Posted: Sun Feb 10, 2019 10:48 pm
by axvers
So, we can't change command (e.g. check_snmp_mem) to resolve this?

if so, we can close this issue, thanks!

Re: Memory False Alert due to cached by os

Posted: Mon Feb 11, 2019 7:43 am
by scottwilkerson
axvers wrote:So, we can't change command (e.g. check_snmp_mem) to resolve this?

if so, we can close this issue, thanks!
Yes that would work.
e.g.

Code: Select all

/usr/local/nagios/libexec/check_snmp_mem.pl -H 192.168.5.182 -C public -2 -w 80,85 -c 90,95
Ram : 67%, Swap : 24% : ; OK