Monitoring software question

Jacob Albretsen jakea at
Wed Oct 9 12:14:17 MDT 2013

On Wednesday, October 09, 2013 02:07:40 PM Dan Egli wrote:
> On October 7, 2013, at 8:37pm, Brian Christiansen wrote:
> > NRPE: Active Check (Nagios initiating the check
> > 
> > NSCA: passive check (the server initiating the check)
> Now I'm lost. I thought Nagios ran on the server? If it does, how does NPRE
> have Nagios initiate the check and NSCA have the server initiate the check.
> That would be the same thing, wouldn't it?

NRPE runs on clients you want to monitor and listens on port 5666 by default.  
You can also limit it to the IP if you have multiple IPs on the box.  Nagios 
plugins to check items is also installed on the client.  The Nagios server can 
then run check_nrpe with a command to check a particular item on the NRPE 
client, which it then contacts and sends the command.  The client has an nrpe 
file which defines what those commands are.  There is some flexibility here which 
can have an impact on security if you enable it (see the docs).  So NRPE on a 
client doesn't do checks by itself, it just does the checks it is told to do 
by the Nagios server.  Each command is one connection from the Nagios server 
to the NRPE client.  So if you run 20 checks, Nagios server connects to NRPE 
client 20 times.

NSCA I have never used, but IIRC, basically the clients runs the checks on its 
own schedule you define and reports the results back to the Nagios server.  The 
Nagios Server is set up to listen for these "reports" from various NSCA 
clients.  The upside of this is NSCA can do 20 checks, and then report back to 
the Nagios server the results in one connection.

More information about the PLUG mailing list