Parallels 18.2 `prl_vm_app` using full 16GB memory footprint of host

I've noticed VMs consuming a substantial amount of memory on our ec2 intel hosts. We have a bug with 18.3 so recently downgraded back to 18.2

Running `for pid in $(ps ax | grep prl_vm_app | grep -v grep | awk '{ print $1 }'); do sudo footprint -p $pid; done` reveals something interesting...

I have `--memsize 1024` yet the VM is still consuming a full 15gb of `MALLOC_NANO`. How Can I prevent this?

```

======================================================================
prl_vm_app [20493]: 64-bit Footprint: 16 GB (4096 bytes per page)
======================================================================

Dirty Clean Reclaimable Regions Category
--- --- --- --- ---
15 GB 0 B 1536 KB 432 MALLOC_NANO
1031 MB 0 B 0 B 66 untagged ("VM_ALLOCATE")
65 MB 0 B 3772 KB 79 MALLOC_TINY
20 MB 0 B 27 MB 160 MALLOC_MEDIUM
12 MB 0 B 0 B 4 MALLOC_LARGE
8244 KB 0 B 0 B 27 MALLOC_SMALL
6774 KB 28 KB 0 B 358 __DATA
824 KB 0 B 0 B 114 stack
736 KB 0 B 0 B 3 IOKit
488 KB 0 B 0 B 3 CoreServices
327 KB 0 B 0 B 116 __DATA_DIRTY
244 KB 0 B 0 B 26 MALLOC metadata
144 KB 0 B 0 B 21 ColorSync
37 KB 0 B 0 B 482 unused dyld shared cache area
36 KB 0 B 0 B 1 dyld private memory
28 KB 0 B 0 B 1 Activity Tracing
17 KB 0 B 0 B 2 __OBJC_RW
8192 B 0 B 0 B 1 os_alloc_once
4096 B 0 B 0 B 2 CoreGraphics
0 B 10 MB 0 B 373 __TEXT
0 B 3080 KB 0 B 10 mapped file
0 B 1076 KB 0 B 13 __LINKEDIT
0 B 0 B 0 B 1 __CTF
0 B 0 B 0 B 1 __FONT_DATA
0 B 0 B 0 B 214 __DATA_CONST
0 B 0 B 0 B 1 __OBJC_RO
--- --- --- --- ---
16 GB 14 MB 32 MB 2513 TOTAL

Auxiliary data:
dirty: -
phys_footprint_peak: 16 GB
phys_footprint: 16 GB
```