New: add Agent UI reverse proxy

• add support for accessing the Agent UI from outside the Agent container by injecting a reverse proxy into the Docker image
• move env variables to .env file and add .env.example

The order in which all the bits work is :

• Dockerfile augments the base DD Agent image by adding a few files, installing nodejs and redefining the entrypoint
• init.sh runs agent launch-gui after 10 seconds (to let the agent to initialize), then sleeps 1 more second (to let the next 3 bullet points below happen)
• that agent launch-gui calls the JS script with the auth URL in argv[2]
• JS script catches argv[2] and writes it to /reverse-proxy.info
• JS script starts the proxy
• init.sh is back from sleep, reads what the JS process wrote inside /reverse-proxy.info and then writes it to the container's stdout
• user sees the URL in the container's stdout and copy/pastes it to their browser

See README.md for instructions on how to use this.