Most of the network monitors I’ve seen for Linux do an admirable job depicting transfers and line speeds, and sometimes even making them look pretty.
Were that the only way to approach the concept, nethogs would never be necessary.
I’ll stop short of saying nethogs is the only network monitor that breaks down transfers by process, mostly because I haven’t seen every program that’s out there. By my calculations, I’ll get to that point around … 2018. 😯
But nethogs does you a service in showing which process is eating bandwidth, and how much.
Other tools I’ve seen aggregate the workload and show how fast or how much is traveling through an interface. That’s cool.
But if your line speed suddenly drops off and you’re not sure what the problem is, nethogs becomes a nifty troubleshooting tool for a search-and-destroy mission.
nethogs keeps things simple, which is always good. Only one or two keystrokes are available while it’s running, and only a few command-line flags to adjust the way it works.
The home page mentions that nethogs, by virtue of relying on /proc, is more or less limited to *nix-ish systems. Don’t look for a Windows version any time soon, I guess.
For me, nethogs is a bit like ncdu: I keep them both installed because occasionally something seems to be taking up too much space, and I want to know what it is. 😈
P.S.: That font is cybercafe. 😉