Nagios Support Forum

Hi all,

I'm network administrator of a small network with 40 switches.

We have Nagios Core 3.5 and some doubts.

I think that Normally, to monitorize a Network with SNMP, the network switches sends traps to the SNMP server.
In some cases, is the SNMP server that asks the switches to check if they have some problem.


NAgios can ask periodically the hosts(network components) to checks the state of some "services", ok.
So, we need to monitorize the network with nagios and I have two doubts:

1º)How do you do to configure NAGIOS to can receive traps of the switches? It's necessary to add NAGIOS SNMP trap interface?

2º)Normally, using NAGIOS, does the network administrators use the traps of the switches or not¿?

Best regards,

Aragorn.

Thanks all¡¡¡

The question is ¿Active checks, passive checks or both to monitoring swithes and routers?

Regards.

The answer to that question I'm afraid is nebulous at best, Nagios tends to do active checking better than passive so wherever feasibly possible for time and effort I opt to check actively. If you are running Cisco or Juniper network equipment than there are plenty of good plugins on http://exchange.nagios.org/ that will give you a large portion of the basic monitoring you need (Usually CPU, RAM, Hardware status and interface status).

I still have a trap service attached to every network device however, which allows our network team to add additional monitoring as they see fit if they want to investigate a specific problem on a specific device. It also allows them to enable anything that might be difficult for me to add otherwise.

Hi and thanks for the reply.

In my case I not work with cisco, I work with 3com.

I think that the active check has some problems:

-You can only monitorize using the plugins that you can found in this page(this it's true?). I'm not sure if the result is the same and I can monitorize all the basically functions.

-You can't have a log of all the traps that the switches send you. In some cases it's usefull.

-The active check uses periodically bandwitch, the traps are send only when i'ts needed...

What do you think about this? Thanks.

Answering your questions in order:
1. You don't have to use the plugins available on exchange, you can write your own plugins... the plugins are just a script or executable of any language that can just as easily be executed from the linux command line as they can by the Nagios application.

2. Nagios will retain logs for active checks, but yes true it is useful to have the logs regardless. Our network team uses Syslog for capturing diagnostic data though rather than SNMP traps. You can go purely SNMP traps if you prefer... it will probably just be a little bit more complex to set up and maintain that's all, Nagios can function aptly either way.

3. They do use more bandwidth (potentially) but it's a drop in the ocean compared to the bandwidth your organization is likely already using... we currently run SNMP queries over 256k links in some areas and it's a negligible amount of traffic.

Unfortunately I've had no experience either monitoring or working with 3com hardware of any variety so as for vendor specific recommendations I can't really offer any advice.