I have a cluster composed of two nodes, let us call them server0 and server1. If a node fails, which is currently the case, LXD commands are not responding, although the containers are running fine.
For example, I launched lxc list 20min ago on server0, while server1 is down, and it still did not return to the prompt.
LXD version is 3.16, installed via snap on Ubuntu 18.04 Bionic.
Documentation is here:
https://lxd.readthedocs.io/en/latest/
Searching on ‘cluster’ will lead you to:
https://lxd.readthedocs.io/en/latest/clustering/
and in this page you will be able to read this:
It is recommended that the number of nodes in the cluster be at least three, so the cluster can survive the loss of at least one node
I missed this one when I build my LXD cluster, or I did not think that it would be such a big deal at the time. Unfortunately, I have only two physical servers. So I guess for now, I will need to deal with the fact that I cannot use the LXD commands if one server is down. At least, containers keep running 
Thank you for your quick answer !