All CTs of host shutted down: "Failed to retrieve PID of executing child process: EOF"

After upgrading to 3.15, one one of my hosts I see that all vms stopped in approx 1-2h tonight.

I the lxd.log I can see many lines like this:
t=2019-07-21T03:32:26+0000 lvl=eror msg=“Failed to retrieve PID of executing child process: EOF”
t=2019-07-21T03:32:27+0000 lvl=eror msg=“Failed to retrieve PID of executing child process: EOF”
t=2019-07-21T03:32:34+0000 lvl=eror msg=“Failed to retrieve PID of executing child process: EOF”
t=2019-07-21T03:32:35+0000 lvl=eror msg=“Failed to retrieve PID of executing child process: EOF”
t=2019-07-21T03:32:39+0000 lvl=eror msg=“Failed to retrieve PID of executing child process: EOF”
t=2019-07-21T03:32:40+0000 lvl=eror msg=“Failed to retrieve PID of executing child process: EOF”
t=2019-07-21T03:32:44+0000 lvl=eror msg=“Failed to retrieve PID of executing child process: EOF”
t=2019-07-21T03:32:45+0000 lvl=eror msg=“Failed to retrieve PID of executing child process: EOF”
t=2019-07-21T03:32:47+0000 lvl=eror msg=“Failed to retrieve PID of executing child process: EOF”
t=2019-07-21T03:32:48+0000 lvl=eror msg=“Failed to retrieve PID of executing child process: EOF”
t=2019-07-21T03:32:49+0000 lvl=eror msg=“Failed to retrieve PID of executing child process: EOF”
t=2019-07-21T03:32:55+0000 lvl=info msg=“Shutting down container” action=shutdown created=2019-07-13T02:37:56+0000 ephemeral=false name=vm646960 project=default timeout=-1s used=2019-07-20T03:35:59+0000
t=2019-07-21T03:34:23+0000 lvl=info msg=“Shutting down container” action=shutdown created=2019-06-22T07:59:25+0000 ephemeral=false name=vm643858 project=default timeout=-1s used=2019-07-20T03:41:53+0000
t=2019-07-21T03:34:48+0000 lvl=info msg=“Stopping container” action=stop created=2018-12-29T06:53:11+0000 ephemeral=false name=vm618292 project=default stateful=false used=2019-07-20T03:35:36+0000
t=2019-07-21T03:34:48+0000 lvl=info msg=“Shut down container” action=shutdown created=2018-12-29T06:53:11+0000 ephemeral=false name=vm618292 project=default timeout=-1s used=2019-07-20T03:35:36+0000
t=2019-07-21T03:35:18+0000 lvl=info msg=“Stopped container” action=stop created=2018-12-29T06:53:11+0000 ephemeral=false name=vm618292 project=default stateful=false used=2019-07-20T03:35:36+0000

Anyone could explain what happened and how to prevent this ?

We would need some information.

I notice you have a few system containers (this is the terminology, and not VMs).
How many, and which container image do you use?

How much RAM does your computer have?

Ok thank you.

For the ram: http://prntscr.com/ohyaeb
There are 59 containers on this host.

Note that i’ve about a hundred hosts running LXD with same configuration, this is the only one I saw this.
Image is debian/stretch on containers (buster on host)