Puzzled by "cgroup_ram_in_use" alert

I’m relatively new to Netdata and still just kicking the tires. Trying it by installing it on a server that is running Proxmox VE which is in turn hosting a number of LXC containers. To be clear, the Netdata agent has been installed only at the host level, not within any container.

I noticed that I started receiving some “cgroup_ram_in_use” warnings. All of them relate to an identified instance of “cgroup_bit-ma-family-ca.mem_usage” where “bit” is the name of one of my LXC containers.

I’m a bit puzzled by what is being measured by the “cgroup_ram_in_use” metric. For example, right now Netdata is reporting a value of around 7.6 GiB. The memory I’ve allocated to the container is 8 GiB. But when I go into the Proxmox webui, Proxmox reports that the related container is using about 500 MB. When I SSH into the container and use the “free” command, that also shows used memory in the container of about 500 MB.

I’m a bit puzzled why there is a discrepancy. Is “cgroup_ram_in_use” for the instance “cgroup_bit-ma-family-ca.mem_usage” measuring something other than the amount of memory being used by the container?

And I guess regardless of why there is a discrepancy, I’d like to understand what should be done to address the warnings I’ve been getting in Netdata. My original thought when I saw it was to increase the memory allocated to that container in Proxmox. But seeing as how both Proxmox and the container itself seem to show rather low usage, not sure if I should.

Any suggestions on how to correctly interpret this metric and/or actions that should be taken to address it, if any, would be most appreciated.