While trying to write a systemd unit for elasticsearch_exporter, I noticed that startup doesn't fail if the process can't bind to the requested port.

A line is logged (level=error ts=2019-05-02T08:28:32.578028364Z caller=main.go:190 msg="http server quit" err="listen tcp 0.0.0.0:9114: bind: address already in use"), but the process just keeps running. In my opinion, this is a bit unexpected behaviour. Instead, the process should exist with a non-zero status code.

Is this intentional, or just an oversight?