I would like to monitor some servers using munin, that are in a different network and are not reachable directly by telnet. I wonder what are the possibilities:

Can I install a central node in the remote network, that collects all data from other servers in this network?

Or would I have to do port-forwarding for each server I want to monitor?

  • 1
    I heavily suggest using graphite instead of munin. That's one of the things it can do and do quite well. With graphite you can have relays that aggregate all the data and relay them back to a master node (or another relay). I.e. exactly what you're looking for. Plus it can scale and do a lot of things that munin cannot. – V13 May 3 '14 at 17:52
  • @V13 thanks! i'll definitely have a look at graphite. – harald May 5 '14 at 8:41

The munin chaps seem to be aware of the need to do this, and they have a project in progress called munin relay, which would allow you to set up a single box at each site that could be seen by the central munin server, and in turn could see all the clients at that site, and which would proxy/relay the munin pull requests from the central server to the (unreachable to the server) clients.

Unfortunately, as the page states, this project isn't working yet. So you can either contribute to the project and make it work, or wait until it's ready, or use one or more of the three other methods listed on that page for contacting the (currently directly-unreachable) clients: tcp proxy, ssh tunnel (or some other vpn, I'd be inclined to look at OpenVPN or IPSec) or ssh:// requests (which will also require port-forwarding).

Or, to summarise the answer, at the moment you've either got to do port-forwarding so the server can reach all the clients directly, or install some kind of site-to-site VPN so the server can reach all the clients directly.


  • Thanks very much and good to hear, that they are working on it! And in the meantime i'll try to get this vpn-stuff to work ... – harald Nov 18 '11 at 12:43

I don't use munin. I use nagios instead, but I think the idea should be the same.

It is much more bandwidth efficient and faster when you install and run the monitoring node in the same area of the monitored servers. You can then simply access the monitoring web interface to check the servers status.

  • mmm ... i understand. the point is, that i have three different networks that i would like to monitor at some central point ... – harald Nov 18 '11 at 12:08
  • Put a collection node in each network then have just those nodes report back to your central instance? – Top__Hat Nov 18 '11 at 12:19
  • Top__Hat: lovely idea, but do you know how to do that with munin? – MadHatter supports Monica Nov 18 '11 at 12:20
  • I think the best option in this case is to install monitoring node in each location and then send a summary of status information to one central location. You need first to find a tool in munin. The other easier option is to install just in one central location if you have a plenty of bandwidth linking all locations. – Khaled Nov 18 '11 at 12:22
  • Khaled, as I said to Top__Hat, that's a lovely idea, but how do you do it in munin? The original poster seems to know that he needs that, so telling him he needs that is no help - tell him how to do it in munin. I use munin quite a lot, and I'm not sure that it can be done, right now. – MadHatter supports Monica Nov 18 '11 at 12:25

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.