You know, anywhere else the above would sound wrong... but here...

I'm probably being thick and making this more difficult than it is, but... here it is.

My company has around 100 locations throughout the country.

At each location we have a LAN with a unique subnet... ex: 10.40.34.0/24 would be a location in South Bend Indiana. The home location is 10.40.1.0/24. They are connected via IPSEC VPN tunnels Regionally, and then from there Nationally.

At each location we have a Windows 2003 R2 or 2008 R2 File and Print Server that acts as that location's "home" server. Regionally we have Exchange 2003 or 2010 servers, so for every 10 locations or so, there's an Exchange server somewhere that their email lives on.

So, at each location we have The Dude running on the server, and we have the key components (printers, wireless access points, etc.) being monitored. We're looking to build a master map.

If we had one dude server at corporate with multiple (local) maps that monitored everything, we'd likely kill the VPN network with monitoring traffic, which is why we implemented the solution we did.

What I'm aiming for is basically what a submap object looks like, but where the map isn't on the local machine.

I see where I can add a remote Dude as an agent, but I can't seem to get the "Submap summary" view (where it says 40 / 0 / 0, for instance).

I can click the lightning bolt in the top left corner, enter the remote dude's IP address, username, and password, and then I can look at the map on that remote dude, but I don't want to be doing this for 100 dude servers every day.

Is there something I'm missing here? Did I not RTFM enough?

Thanks, in advance!

We have a distributed Dude monitoring system - similar to what you are trying to do but not identical. We have set up different Dude agents in different segments of our network (firewalled from eachother) and use the "local" agent to poll the different segments. Firewall rules allow the Dude Master and Agent to communicate.

The Master Dude is the one that we connect to with the Dude Client. I think that it holds all of the map information. On the Master, under the "Server Configuration -> Agents" tab we added each segment Dude Agent.

Then when we add a device to a map we select which agent is used to poll that device (on the device's "General" tab)

I can't say for sure if the Agent actually originates and manages the polling - reporting status back to the master, or if the agent just proxys the master polling. But this is the way to set up remote agents, you will have do some testing to work out if it will suit your needs.

Alternatively, you could set up each site's Dude server to be it's own system, write a script that would update a central web-page whenever a device at the site changes state and then have links on the central web-page that will connect you to the site's Dude server.

Again, I might be reading too much into this, but...

My impression is that whereever the device is set up, that's what issues the commands. The command could be as simple as "hey you, agent206 - go check and see if X device is online and report back to me". But that device is the "thing" that issues the commands.

So if I have a master Dude server set up at our corporate office that has all the thousands of devices set up on it, that means it's issuing thousands of commands to the various Dude agents over the WAN... which creates some serious network traffic overhead, because when the agent gets a response it has to send that response back to the master server. So that's 2 sets of traffic (command, then response) that traverses the WAN. No one wants their network monitoring traffic to be a significant percentage of their WAN traffic.

So, if I read between the lines (aka lack of multiple responses), the conclusion I come to is you're correct, Geoff.

So here's a feature suggestion to the devs that create The Dude (thanks, guys!) - I would suggest that instead of the above, Agent servers need to be able to act independently to query their respective devices. Then on the master console or even regional master consoles, you could right click on the map and select "Add Dude Agent...", and you'd point it to the IP address, username, password, and map that you want stats off of. After doing so (and assuming you do it all correctly), you'd get an icon that says "[Location] [MapName] (51 / 3 / 1)" and the icon would match the status.

Possible?

Again, thanks in advance, all!

I have not used agents in the dude but when you add a device you can select the Agent. I would assume that if you specify the agent the device is polled from the agent and only changes/outages are communicated to the master server. As far as bandwidth, having 1000 services set at 2 minute polling and that requires about 250kbps. I think you are safe to build as many agents as you need then if bandwidth is an issue just up your polling time.