Hi @adamsmithad ! I took the liberty of moving this post to a more related category so that hopefully it will get more visibility.
Your question seems specific - about ethernet adapters and cloud environments, but in practice from what I’ve seen it doesn’t quite go that way. In a lot of cloud environments things are virtualized to the point where you have network interfaces at a guest OS layer, but you don’t really have any etherenet adapters per se.
Trying to give you the closest possible answer, we’d use something like node_exporter or grafana agent to monitor a linux host. That would include network statistics (such as bytes transmitted/received) which would come from the network adapter of the host, but isn’t really all that specific to ethernet adapters.
To get further down into it, things like node_exporter and agent have “collectors” which gather prometheus metrics from lots of different hardware devices. On this page you can see all the various collectors that are available: GitHub - prometheus/node_exporter: Exporter for machine metrics
Some are enabled by default, others you have to enable. These individual things get way down into considerations like what kinds of hardware are you supporting, and what OS are you running.
To make matters more complex, the specific type of hardware or ethernet adapter will expose different kinds of metrics depending on what it does.
How can we ensure that our monitoring setup is scalable and can handle large amounts of data from Ethernet adapters in the cloud?
Follow proper grafana agent config details and report those metrics to an auto-scaling cloud instance. 
What are some best practices for monitoring Ethernet adapters in cloud-based infrastructures using Grafana?
That is a very big question I wouldn’t attempt a simple answer to, other than saying “first start out with a sense of what conditions you want to monitor, what can go wrong, and how you wish to respond”. This guides a lot of the “how” of monitoring.
