From reading this forum, I found out about /proc/io_status. Presumably if I continually run out of tokens I'll get throttled, as apparently happened to the guy that wrote the post I saw.

Is there anything else I should keep an eye on from within the VM? I take a look at my resource usage in the web interface when I have occasion to log in, but if there's something within the VM itself I can watch automatically (eg nagios) then I stand a much better chance of catching it before it becomes an issue for anyone else.

You can watch /proc/io_status , maybe have your system email you if your tokens drop below the 50% mark or something.

Yeah, I do that. I was more thinking about indicators for CPU or network usage though.

Oh. Well try this one then:
https://www.linode.com/members/graphs/index.cfm

Check out /proc/stat

Specifically the 'swap' line. The first number is the number of swaps in and the second is swaps out. If either of these is growing too quickly you're swap-thrashing and we will all hate you.

The 'page' line is for page in/out but there isn't much you can do about that as far as I know.

If you want a pretty, human readable version to look at, grab the 'procinfo' program.

--James

FYI: My linode has been up 3 days and I have 80 swap ins and 130 swap outs. Those are good numbers. Very little swapping going on.

I'm curious about what I can check to stay in good graces with my neighbours so I wanted to check this out. I see no swap line:

harleypig@harleypig ~ $ cat /proc/stat
cpu 528273 40 239602 229393529 406132 0 125 0
cpu0 528273 40 239602 229393529 406132 0 125 0
intr 282928833 279326787 0 1 0 2972324 629721 0 0 0 0 0 0 0 0
ctxt 18861030
btime 1169363940
processes 797113
procs_running 1
procs_blocked 0
harleypig@harleypig ~ $