Comment 9 for bug 1804149

Not sure if I understand the overall situation correctly, but:

In the bug description I read that the issue occurs with 4.15.0-36-generic,
and in comment 4 its also mentioned that it happens with 4.15.0-36-generic and 4.15.0-23-generic
and again in comment 6 also on 4.15.0-15-generic (I assume all on 16.04).

But in comment 8 (according to LP 1782085) the issue doesn't happen (at least for the last 47 days) on a kernel 4.15.0-36-generic (on that particular system) - but this is the same kernel level like the one mentioned in the bug description?!
Either there is something wrong with the mentioned kernel levels, or it doesn't depend on the kernel, but for example on the infrastructure?!

LP 1782085 describes a situation on an LPAR system, but LP 1804149 on a zVM 7.1 guest system.
Are there other differences between both systems?
Can it have something todo with the underlying hypervisor?