I used to have a Nagios, but after years of continuous uptime (except for planned maintenance) i felt it was not worth it. If your tech stack is simple enough and runs on VPS'es (whose physical availability is responsibility of your hoster), there isn't much that could happen.
If i were to setup metrics, the first thing i would go for is the pressure stall information.
If i were to setup metrics, the first thing i would go for is the pressure stall information.