The Debian init system general resolution returns

Are you actually, legitimately suggesting that a fully working init repalces the need for Nagios, or just trying to score rhetorical points? In any event you could certainly use a more complex service check like what you run under nagios and then communicate with systemd either with systemctl or over the api to shutdown or restart problematic services. You probably could even make a watchdog unit that is kicked off by a timer and runs the complex availability check when the service you are monitoring is enabled, maybe that becomes a new standard for how high availability services are monitored.