Q: Monitoring ISP routers with auto discovery?

[jimmy_1969]

Hi,

I have a use-case for Nagios that I can't get my head around.

I want to monitor the route between by my firewall and the ISP's Point-Of-Presence to public internet. In this case there are three router hops b/w my firewall and the PoP. But the actual path changes dynamically and so does the IP addresses.

Is there a way of dynamically discovering IP addresses, e g based on traceroute, and adjust the Nagios config based on this?

Thanks in advance.

Best Regards

//Jimmy

[jolson]

Would there be any issue with running a manual traceroute from your Nagios box to the ISP router in question, and then generating a new host for each discovered router manually?

[jimmy_1969]

Hi Jolson,

Well that is what I have to resort to for now. Traceroute & update Nagios ISP router IPs. Next time my DHCP lease expires I might find myself on a new subnet and with a completely new set of router IP addresses on the path to the ISP's PoP. And a new Nagios config update.

This is a bad solution as every time I get Nagios alarms related to my ISP's equipment the first thing I have to do is validating that they are not false positives. For a proper monitoring solution that question should not have to be part of the equation.

And, yes, I could start mapping out all ISP subnet- and routing combinations but that would be messy to maintain over time. So I would be interested how to make a dynamic Nagios configuration, e g based on parsing the router IP addresses of traceroute, so I can direct the monitoring effort only to the active path nodes.

//Jimmy

[jdalrymple]

I think it's really awkward that your ISP is assigning a different router to you when your dynamic address changes. If my ISP did that I'd change ISPs. If that's not an option, perhaps the easy solution is to see if your ISP will give you static addresses? My ISP gives me a whole /29 netblock with my cheapo commercial account and I think this is common with most business Internet packages.

If that's still not an option you're stuck with crafting your own solution I guess. Our monitoring solution fully abstracts how to monitor from what to monitor. Generally speaking if there is something out there worth monitoring then a plugin has been written for it already and is posted on the Exchange. If your use case is so narrow that nobody else has already written a plugin for you, then this is where you get to do some legwork yourself. You can write your own C program, perl, Python, Ruby, Shell, PHP, whatever script, or whatever else you want to monitor whatever you want. This is where we tend to think our monitoring solution is "proper" as the monitoring options are truly unlimited. If doing a bit of work to achieve the customization you need for the application you have isn't in your scope, we do have consulting services available in-house or there are hundreds of Nagios consultants worldwide that would be happy to write a plugin to suit your application for you.

I'd be surprised if you found any pre-built route hop monitoring solution that handled route changes on DHCP release/renew properly. That's just not a very common practice.