Gateway events for IPv6 affect IPv4 services and vice versa
When a gateway of a specific type has an event, all services on the interface are restarted, rather than just those relevant to the type of gateway that had an event.
For example, if you have dual stack IPv4 and IPv6, and separate OpenVPN servers on IPv4 and IPv6, then the IPv6 gateway going down will restart both, not just the IPv6 instance. This is unnecessarily disruptive to the services that did not have an outage.
Ideally, a service should only be restarted if it is dual stack or if it's using the same IP type as the gateway that had an outage.