Memory use of containers and tmpfs

[luison]

We are trying to debug some memory issues with a container and best practices to monitor

Within the PVE interface I see
Memory usage
11.59% (949.30 MiB of 8.00 GiB

inside the container:
cat /proc/meminfo

MemTotal: 8388608 kB
MemFree: 2967096 kB
MemAvailable: 7401044 kB
Buffers: 0 kB
Cached: 4433948 kB
SwapCached: 0 kB
Active: 3422264 kB
Inactive: 1725752 kB
Active(anon): 797748 kB
Inactive(anon): 817052 kB
Active(file): 2624516 kB
Inactive(file): 908700 kB
Unevictable: 0 kB
Mlocked: 17976 kB
SwapTotal: 8386552 kB
SwapFree: 8384948 kB

There should not be any swap

And from the host if I issue:
lxc-info 1111
Name: 1111
State: RUNNING
PID: 6607
IP: 10.0.0.111
CPU use: 27060.46 seconds
BlkIO use: 0 bytes
Memory use: 5.18 GiB
KMem use: 267.18 MiB
Link: veth1111i0
TX bytes: 965.51 MiB
RX bytes: 664.02 MiB
Total bytes: 1.59 GiB

Could this be related from the fact that the container has 2 tmps mounts for tmp and cache?
Can someone confirm if that tmpfs is reduced from the 8Gb assigned to the container or from free memory on the host?
Are there any commands on the host to monitor all tmpfs assigned space to containers??

Thanks.

 

[luison]

Currently upgrading a server to PVE6 and trying again to understand where is tmpfs taken from on a container.

Is that tmpfs assignment coming from the total allocation of the container or from the host?