LXC wrong stats reported

General
1

I have latest Netdata installed at a LXC container and get now apperently stats from the host instead of the LXC container.
The LXC Container (named ansible) is hosted at a Proxmox installation and the ceph is used as container storage.

API info

It looks like it mixes hosts stats and container stats, I am just installed it with no further configuration, so I am unsure if this is some behavior I could adjust or is not supported ?

root@ansible:~# df -h
Filesystem      Size  Used Avail Use% Mounted on
/dev/rbd13       15G  3.0G   11G  22% /
none            492K  4.0K  488K   1% /dev
tmpfs            32G  184K   32G   1% /dev/shm
tmpfs            13G  148K   13G   1% /run
tmpfs           5.0M     0  5.0M   0% /run/lock
tmpfs           6.3G     0  6.3G   0% /run/user/0

image
image1082×936 55.6 KB

Here it reports some data I have no idea from where it comes from.
The 54,76GiB the total 10.24TiB are unreasonable and can’t get find at the host either (or some combination from some values that are unknown to me)
On the other hand the 3GiB free storage is correct.

At the memory tab it looks okay besides the total committed memory that is completely off the charts and likewise a value from the barebone host.

image
image1050×625 35.9 KB

2

Hi. These are LXCFS limitations:

LXCFS will take care that the information provided by crucial files in procfs such as:

/proc/cpuinfo
/proc/diskstats
/proc/meminfo
/proc/stat
/proc/swaps
/proc/uptime
/proc/slabinfo
/sys/devices/system/cpu/online

are container aware such that the values displayed (e.g. in /proc/uptime) really reflect how long the container is running and not how long the host is running.

It means that other procfs files provide the host data. This is why you get mixed container/host metrics when installing Netdata inside an LXC container.

3

I had a look at /proc/diskstats and they are exactly the same at host and the lxc container :confused:
So the disk stats are completely useless, not sure if this is a proxmox (using the latest proxmox, lxcfs 6.0) problem or a general lxc problem.

For example /proc/stat looks suspicious too, so cpu count looks okay but some values are basically the same at host/container

ctxt ...
btime ...
processes ...
procs_running ...
procs_blocked ...
softirq ...

The other stats are likely similar.

4

Yeah, I have 2 Proxmox 8.1 boxes and see the same issue for /proc/diskstats.