How to fix [ServiceUnavailable: WebSocket connection failure] error on neo4j-enterprise AMI deployed on EC2?

Sorry you're running into so many issues, but I can see from your post that there are a number of things going wrong. Here's a rundown of what they are and what you can do to fix them.

The "service unavailable" error is caused by the fact that a newly launched amazon VM does not have an SSL certificate. This occurs because your browser is trying to connect to both ports 7473 (HTTPS) and 7687 (bolt) and many browser won't do this if the SSL cert is untrusted, without you first trusting it.
See the second issue in this post:

That post has a list of resolutions, including getting your own certificates.

Also, if you're modifying /home/ubuntu/neo4j.conf, this is the wrong file to modify. There is documentation here about how Neo4j Cloud VMs work, and the file you should be modifying is /etc/neo4j/neo4j.template

Additionally - the command used to restart the system service is systemctl restart neo4j not service neo4j.

Finally -- changing the line with dbms.connector.bolt.address will not resolve the error you're encountering. You need to either have your browser trust the SSL cert, or get a trusted SSL certificate.