A lot of folks have asked to fix various cumbersome parts of the node import / discovery process
I suggest this as both a stop-gap method and as a way to allow customer to create a node import list using whatever means they desire.
In many cases this will be a script, or the results of many script processes, possibly run in parallel from different locations.
It also wouldn't matter how long it took to create the list, and it would likely offload a lot of NPM work vs. each list item.
If we had a CSV template that allowed us to import nodes, including the interfaces, volumes, etc. that we did/didn't want, it would simplify and speed up the entire process.
Picture a column for each type of node property we wanted to see, or perhaps even a keyword (like ANY, FIXED, or VIRT) to indicate everything in (say) the 'Volumes' column that we cared about.
This could also finally let us create templates of node groups for repeated (batch) import or update, with NPM handling any new or changed node information.
It puts node updates into the hands of the users, giving us the freedom to maintain these node list any way we wish.
NPM could be configured to react in different ways on errors: don't import, mark as unmanaged, etc.
I'd use this feature a lot.