Workarounds for un-containerized resource limits visible in containers

Typical cases: `free` and `top` commands.

Some references:

- https://fabiokung.com/2014/03/13/memory-inside-linux-containers/
- https://ops.tips/blog/why-top-inside-container-wrong-memory/
Currently there is no way to change the information exposed by `/proc/meminfo` and `sysinfo()` system call without patching the Linux kernel manually.

Potential solutions:

- Wait until such well-known tools are updated to used `libmymem` which takes cgroup/rlimit into account when calculating available memory.
- Update our jail to override the content of `/proc/meminfo` and the result of `sysinfo()` system call.\* Currently overriding an arbitrary file content via jail is very tricky, since we need to emulate file I/O syscalls to work upon a custom-built buffer.

┆Issue is synchronized with this [Asana task](https://app.asana.com/0/1159751085623729/1159757054588980) by [Unito](https://www.unito.io/learn-more)

JIRA Issue: BA-320

Start by tracing the jail's handling of /proc/meminfo and sysinfo(), then compare free and top output with the container's cgroup and rlimit values. The issue proposes either libmymem-aware tools or jail-level overrides for these interfaces; determine which path is supported and validate the resulting command output.

Written by the indexing model from the issue text.