[dralley]

As someone pointed out elsewhere in the thread, so can systemd, your service did what it was configured to do. If you want the service to stay up when netdev goes down, then don't tell systemd that netdev is a hard requirement for your service.

[egberts1]

Its never a good systemd-networkd default to be killing daemon whenever a netdev goes “cricket”.

[dralley]

But that's just it, to say this just demonstrates you don't understand the systemd service model.

There is no "system default" in the loop here at all. The service told systemd that it cannot operate without netdev, and systemd behaved accordingly. If the config was written more appropriately, it would have behaved accordingly. The "default" is what you had written in the unit file.