fix(healthcheck): use address configured if not empty

[qdm12]

Problem is I have FB_ADDRESS=10.0.0.28 (my internal bridged network), and the healthcheck always fail with localhost:

/ # curl -f http://localhost:80/health
curl: (7) Failed to connect to localhost port 80 after 0 ms: Couldn't connect to server

Changing localhost to the value of $FB_ADDRESS fixes it:

/ # curl -f http://10.0.0.28:80/health
{"status":"OK"}

The proposed changes pick the address using the following:

1. Use $FB_ADDRESS content if not empty;
2. Fallback on "address" value in /.filebrowser.json
3. Fallback on localhost

[Guiorgy]

NGL, I'm not very versed in networking, but shouldn't localhost always loop back?

[aseg187]

this change causes unhealthy status with error message = curl: (3) URL rejected: Bad hostname revert back to old file
#!/bin/sh
PORT=${FB_PORT:-$(jq .port /.filebrowser.json)}
curl -f http://localhost:$PORT/health || exit 1

[qdm12]

Indeed, my apologies. It looks like this fixes it:
#3085 (comment)

Should I open another PR or someone did already?