Does that mean that each vserver keeps its own statistics? Because vnstat also parses that output: user@vserver:~$ strace -e trace=file vnstat -tr 2 execve("/usr/bin/vnstat", ["vnstat", "-tr", "2"], [/* 12 vars */]) = 0 access("/etc/ld.so.nohwcap", F_OK) = -1 ENOENT (No such file or directory) access("/etc/ld.so.preload", R_OK) = -1 ENOENT (No such file or directory) open("/etc/ld.so.cache", O_RDONLY) = 3 access("/etc/ld.so.nohwcap", F_OK) = -1 ENOENT (No such file or directory) open("/lib/x86_64-linux-gnu/libm.so.6", O_RDONLY) = 3 access("/etc/ld.so.nohwcap", F_OK) = -1 ENOENT (No such file or directory) open("/lib/x86_64-linux-gnu/libc.so.6", O_RDONLY) = 3 open("/home/user/.vnstatrc", O_RDONLY) = -1 ENOENT (No such file or directory) open("/etc/vnstat.conf", O_RDONLY) = 3 open("/proc/net/dev", O_RDONLY) = 3 Sampling eth0 (2 seconds average)...open("/proc/net/dev", O_RDONLY) = 3 42 packets sampled in 2 seconds Traffic average for eth0 rx 8 kbit/s 10 packets/s tx 16 kbit/s 10 packets/s user@vserver:~$ If I can just use vnstat that's fine by me, but I was under the impression that these statistics were shared? On 2014-12-23 16:08, Jean Weisbuch wrote: > You can read how much data has transited on each network interfaces > inside a guest by reading : /proc/net/dev > You just have to take the value at first then wait some time, take > the new value, substract the first one to the second then divide it by > how long in seconds you waited between the two checks. > > If you want to check from the host, here is a script i made : > http://paste.debian.net/137926/ > > Le 24/12/2014 00:45, Laurens Vets a écrit : >> Hi list, >> >> Is it possible to monitor the bandwidth of 1 particular guest from >> inside that guest? >> >> Basically, I want to run vnstat inside a guest monitoring only the >> traffic that guest generates. Is that possible somehow?