I have containers that are sitting behind an Nginx reverse proxy. In order to monitor them via Uptime Kuma on a different physical host, should I:
a) Point Uptime Kuma at the FQDN of the container that goes through NPM?
Or:
b) Point Uptime Kuma at the [internal IP]:[port] of the container?
My concern with option A is I’m not really sure if it’s heartbeating off the target container, or simply the NPM. Meaning, the container could be down but NPM is up therefore Uptime Kuma thinks the container is up.
Isn’t the whole point of a reverse proxy to use FQDN instead of IP:port? Why would you monitor IP:port in that case? You want to know if you’re unable to access your service, if you’re monitoring IP:port and your reverse proxy shits the bed, then Uptime Kuma will happily report everything up and running. If you’re monitoring FQDN and either your service dies or your reverse proxy dies, UK will notify you your service is unreachable, granted if your RP dies UK will alert you everything is down, which is also more helpful as you can narrow down the issue.
