Comment 31 for bug 1871538

That's interesting actually. I thought this was an interaction between reloading dbus-daemon and daemon-reloading systems at the same time. But in this cloud case, I don't think dbus is being reloaded, is it?

It's going to be next to impossible for me to look into this in detail without being able to reproduce it or getting a bit more debug information.

Maybe you could crank up the systemd log level and try to reproduce this?