needrestart currently defaults to restarting virtlogd when it detects
that the service needs a restart.
According to the man page and systemd unit, a reload operation will
re-exec the virtlogd executable without losing virtual machine logs.
Please ship /etc/needrestart/restart.d/virtlogd.service and
/etc/needrestart/restart.d/virtlog scripts which should call 'systemctl
reload virtlogd.service' and '/etc/init.d/virtlogd reload',
respectively, so that needrestart doesn't just restart the service on
its own.
See /etc/needrestart/restart.d/README.needrestart for more information.