[Bug, CI, Scripts, Compose] - Default compose settings for repo domain

[JoshuaSBrown]

Description

Problems with allocation requests hanging in the task queue if the repo tcp address does not use an ip address when filling out the repo form.

Trying to use a domain name in the repo form or the Compose container name does not seem to work correctly. I.e.

tcp://localhost:9000
tcp://<domain_name>:9000
tcp://<compose container name>:9000

Using the actual IP address of the repository instance will allow communication to between core and repo to work correctly

The reason for this is because all the traffic gets routed to the web server if a domain name is used.

Steps to Replicate

Additional context

Page occured:

Expected behavior

Artifacts (if appropriate):

[JoshuaSBrown]

Adding the internal-network to the repo container in the docker compose file allows us to specify the address tcp://datafed-repo:9000 and work with the allocation creation request.

Also works with the domain name.

     ports:
       - 9000:9000 # Communication core server
+    networks:
+      - datafed-internal

[JoshuaSBrown]

localhost does not work - I think because the core service assumes localhost refers to its own internal container.