Very slow copy to remote host

Both host on the same network, connected with 10G fibre. In the past we have been backing containers at about 40MB/s, but this has dropped to less than 8MB/s for some reason.

rob@farmer:~$ lxc info gitlab
Name: gitlab
Location: none
Remote: unix://
Architecture: x86_64
Created: 2018/10/02 22:23 UTC
Status: Running
Type: persistent
Profiles: default
Pid: 4152
  eth0:	inet	vethGF4H08
  eth0:	inet6	fe80::216:3eff:fe7e:a4be	vethGF4H08
  lo:	inet
  lo:	inet6	::1
  Processes: 550
  Disk usage:
    root: 2.92GB
  CPU usage:
    CPU usage (in seconds): 107
  Memory usage:
    Memory (current): 2.19GB
    Memory (peak): 2.19GB
  Network usage:
      Bytes received: 702.87kB
      Bytes sent: 15.35kB
      Packets received: 1812
      Packets sent: 203
      Bytes received: 1.94MB
      Bytes sent: 1.94MB
      Packets received: 1502
      Packets sent: 1502
  gitlab-test (taken at 2018/10/10 01:16 UTC) (stateless)
  backup (taken at 2018/10/18 02:01 UTC) (stateless)
  current (taken at 2018/10/18 02:02 UTC) (stateless)
rob@farmer:~$ lxc copy gitlab/backup hunter:gitlab-18-10-18
Transferring container: gitlab-18-10-18: 291.11MB (6.00MB/s) 

A lot of the time copy just hangs without any Transferring container information.

Both server are running 18.04

rob@farmer:~$ lxc version
Client version: 3.6
Server version: 3.6

rob@hunter:~$ lxc version
Client version: 3.6
Server version: 3.6

Happy to supply any details needed to resolve.
Any help or direction would be appreciated.

What storage backend is that?

ZFS at both ends.