cgroup memory limits not shown when setting up a cluster/in prometheus

As part of testing MB-52686 we found a couple of places in ns_server where cgroup limits don't appear to be respected, although it's unclear from the original ticket and document whether they should be.

1. When a new cluster is created the cgroup limit is not shown as the maximum amount of memory that can be provisioned
2. The sys_mem_total and sys_cpu_cores_available prometheus stats don't respect the cgroup limit

The testing was done on Docker for Windows using the ubuntu:20.04 image and installing the latest neo couchbase server deb package (as of 1st July) manually.