Nagios Support Forum

My client has expressed need for virtually immediate alerts for CPU, failed disk, disk full, RAM, etc. and I'm wondering how to best meet the need.

First thoughts are a hybrid solution of active checks with check_by_ssh along with SNMP.

Can you please assist in designing most accurate and immediate monitoring system for both windows and linux?

Immediate alerts are usually passive checks that are monitored by the monitored server and sent to Nagios (i.e. SNMP Traps)
Personally I think that for the checks you've mentioned it's possible to use regular SNMP monitoring with a check interval that is less than 1 minute.

So just configure pure SNMP checks, and do without agent?

Can NSCA be used just as effectively as SNMP?

asmgiadmin wrote:Can NSCA be used just as effectively as SNMP?

I would assume so. You may want to decrease the freshness check interval a bit so that passive checks are reported by the UI faster.
http://nagios.sourceforge.net/docs/3_0/freshness.html

It depends on what you are looking to check with NSCA. NSCA offers passive checks so once a change in what it is monitoring occurs it will send a update to the Nagios server.

Please see our already prepared documentation:

http://nagios.sourceforge.net/download/ ... _Setup.pdf

http://assets.nagios.com/downloads/nagi ... ith_XI.pdf

Bear in mind if you do decide to use NSCA please use the stable version and not the current version as there are a few bugs being worked out.

Consider monitoring solution is NSCA and CPU, RAM, and Disk checks will be performed.

Is it correct that each check above will require script that will then pass value to NSCA to relay to server?

In addition to abrist's links, for a general overview of how NSCA functions you can visit the NSClient++ site and view their short guide:

http://www.nsclient.org/nscp/wiki/doc/usage/nagios/nsca

NSCA can use plugins, but in the case of standard windows checks it will use a loaded module from the NSClient++ service.

I'm asking more in regards to Linux monitors.

If CPU, RAM, and disk are to be monitored on server, will script need to be created to call plugins and pass result to NSCA?

Most people monitor Linux system actively with NRPE or our XI Linux agent