I am not sure if the bellow metrics are already available in the Linux version (I have only used the Windows version for now), but I think it would help a lot the node operators to understand if the nodes are healthy or not for a particular reason.
- Average request size (it can also have max/min but this is a suggestion).
- Response time per request (along with time to service the full request, etc, just to give the operator some idea of how bad/good the node is able to service the network).
- Number of requests being processed (real-time or near)
- Number of requests per day (add to a graph if possible)
These are just examples to help a node operator understand if the node is operating in a healthy way. Would also be interesting considerations to add to a node reputation if not already there.