Tips/Tools to diagnose network usage (high ping or ping spikes)

I've been wondering for a while, how do you guy usually diagnose network problems in a server?
Take into consideration most people run their servers in a panel (currently including myself) and most likely don't have access to the host OS.
Also, this is assuming the ping is caused by the server rather than the host/network.

For performance diagnostics (tps/mspt/ram) our standard is spark, yes
But how do you do the same but for network? spark-viewer only has a small section dedicated for network, but I find it somewhat lacking, mainly because when testing for small playercounts, all numbers look "fine" (green).
Is there another tool that you use that I might be missing here to do a slightly more detailed network analysis? I'm aware tracking network usage per-plugin will most likely be impossible, but maybe a graphic would be helpful.
Or maybe a table/graph/standard/indicator to follow to identify how many bytes/s or packets/s are "too much" for each player count?

For example, when I test random plugins I find, and one of them makes my ping rise to 200 or even 2k (even in localhost, which should be around ~0ms), I usually track the culprit by taking a memory profile in all threads and see which one makes the most allocations, specifically packet creation.
While this has worked fine for me so far, it is a faulted approach since I have to assume the packets are not being re-used