rockhost, I was just directed here by one of my staff and wanted to clarify the reasoning behind how this works.
First, just a note that Centova Cast doesn't bind to port 80 at all unless you enable the port 80 proxy. It's actually the port 80 proxy that's returning that "Nothing here" message.
The reason we don't have it redirect to port 2199 by default is because that would only work for port 80 proxies installed on the same server as the web interface. When you run out of capacity and add another server to your hosting service, your second server runs its own port 80 proxy, but still uses the web interface from your first server. (See the manual for details.)
So to draw you a bit of a picture:
Server A: 1.2.3.4, running web interface on 1.2.3.4:2199 and port-80-proxy on 1.2.3.4:80
Server B: 4.3.2.1, running port-80-proxy on 4.3.2.1:80, and nothing on 4.3.2.1:2199
If server A's port 80 proxy says "aha, I'm running on 1.2.3.4:80 -- let's redirect to 1.2.3.4:2199" that'll work fine.
If server B's port 80 proxy tries the same thing, it'll redirect to 4.3.2.1:2199 and the user will get..... well, nothing really, just a "server connection failed" message in his browser, because there is nothing listening on 4.3.2.1 port 2199.
Granted, the IP of the web interface is passed to the slave server during installation so it would be possible to script some configuration file modifications at that time to tell the proxy to redirect to the web server's IP address, but that seemed a relatively low-priority feature which required more fiddling than it was worth when we were in a time-crunch to get v3 released.
We may revisit this in future (it's been added to our tracker), but it's not as simple as just adding a simple hardcoded rewrite rule to nginx or a line of JavaScript to the default index.html which is why it's not there now.