Issue Details

health_port setting is ignored even if we don't define health_upstream.
This behavior occured after upgrading from version 2.10.2 to 2.11.1.
We use Garage S3 server and it was working properly with this example config until we upgraded:
https://garagehq.deuxfleurs.fr/documentation/cookbook/reverse-proxy/

s3.garage.tld, *.s3.garage.tld {
	reverse_proxy localhost:3900 192.168.1.2:3900 example.tld:3900 {
        health_uri       /health
        health_port      3903
        #health_interval 15s
        #health_timeout  5s
    }
}

Assistance Disclosure

No AI used.

If AI was used, describe the extent to which it was used.