How do I monitor latency between two nodes in our environment?

Unless I'm getting the wrong end of your question, then so long as you have the specific interface(s) that connect these two nodes together, then you can go to the specific interface summary pages for each and see traffic, etc...

Alternatively you could set up a specific Perf Analysis project, add the distinct interfaces in and whatever metrics you want.

You've got some options

If they send traffic to each other and you've got agents, you can enable packet scanning stuff for this and get auto-mapping, it's heavy though.

You can promote an agent to a mini-poller and monitor one box from the other like that if you're on licensing that allows it
If you've got SAM you could put a little script on to check the other one

If one or more of these devices is cisco you might be able to use IPSLA with VNQM

Probably a few other choices around

Thank you very much for the response, its much appreciated.

The trouble is that two given sites don't have a link which is exclusive to those two sites. So lets say that we have a regional colo in NYC, which hosts resources for a number of other US sites. If I look at the network utilisation of the colo site in NYC, I'm seeing the collective traffic of all sites which access those resources.

So I'm not sure I can gauge latency between the colo NYC site and say denver. As the link isn't exclusive.

Hope that I am making some sense, I'm a solarwinds person rather than a network engineer

 Remote Ping using any Network Device Maybe this is of use?

 kpmarcin made a great video tutorial on how to set this up.

Alternatively check this out  Remote TCP Check 

Ahh understood and yes I do follow. In which case you'll only ever get a "sense" of the latency, etc by monitoring specific links.

If you have Netflow and the relevant nodes are configured to send that data to Solarwinds, then you could use that to extract the conversation(s). Other than that, then Adam's suggestion is a potential option depending on the node(s).

I see some great replies, what are the nodes that you are referring to here ? Since this is a question asked in NPM forum I am assuming its Network Infrastructure ? Like Router, Switch etc or are these Windows/Linux boxes ?

__PRESENT__PRESENT__PRESENT__PRESENT__PRESENT

To be fair I put it into NPM as it seemed to be the best place for the question, however, a latency between between 2 domain controllers would probably server me fine as well.