Why Confluence is so sensitive to FQDN ?

[padauk]

To setup Confluence server with reverse proxy, one needs to set consistent FQDN to all of Server Base URL, proxyName of server.xml and reverse proxy setting of Apache/Nginx. This is not simple (†1). 

What is the reason for all of these to be consistent ? There are many web applications which works without the care of FQDN part of the URL.

(†1) Especially when the application server is runing in private network with external access with port mapping, host naming of private network is different from the external (public) network which makes Confluence setup complicated.

[AnnWorley]

I understand your naming scheme is different internally than on your public network so you would like to be able to access Confluence from either one.

Confluence does need to be accessed on the base URL to work properly. Internal links are built using the base URL and also the plugin manager depends on the base URL. For more insight into why a single base URL is required, please see: Various Issues Caused when Server Base URL Does Not Match the URL Used to Access Confluence

We do have a suggestion open to allow Confluence to be accessed on more than one URL:

Support accessing Confluence from alias that is different to the server base URL

You may vote and/or comment on the ticket to emphasize your use case for the development team.