Problem Statement

I am running a Flask-based API within a Docker container that is exposed on the port 5000, however no requests are even getting to the API.

To reproduce

1. Get jupyter/all-spark-notebook docker image
2. Within the Docker container, get any Flask-based API, for reproducibility purposes try https://github.com/miguelgrinberg/oreilly-flask-apis-video (for understanding a very simple Python-based API switch to v0.1 in releases)
3. Run the Docker using the standard command as described here (https://github.com/jupyter/docker-stacks/tree/master/all-spark-notebook). Forget about the notebook part, and only use -p 5000:5000 to get the API preferred port to run. Don't use HTTPS, but do install ssh on the container. (So unless you are using a custom Dockerfile like me, you will need to enable SUDO)
4. Try to make any request to API (e.g. http GET http://localhost:5000/customers/1) --> this fails and gets different error messages depending
   • whether it is Python based urlib request (IOError: ('http protocol error', 0, 'got a bad status line', None))
   • or HTTPie based (http: error: ConnectionError: HTTPConnectionPool(host='localhost', port=5000): Max retries exceeded with url: /filterReplies/aaaaa/aaaa (Caused by <class 'http.client.RemoteDisconnected'>: Remote end closed connection without response))
5. Check logs, where no record of any API request

Notes

• I am running on a linux machine so no VM necessary.
• There is a difference in making a wget http://localhost:5000 (gets Connecting to localhost (localhost)|127.0.0.1|:5000... connected. HTTP request sent, awaiting response... No data received. Retrying.) and wget http://localhost:1234 (gets Resolving localhost (localhost)... 127.0.0.1 Connecting to localhost (localhost)|127.0.0.1|:8888... failed: Connection refused.) so the port is open on some level, but there just does not seem to be anything waiting on the other side?
• The original purpose of the image still works--i.e. if I specify both ports to be forwarded, I can still have a browser-based access to the Jupyter notebooks within the docker container. So some port forwarding on this machine works.

Binding to 0.0.0.0 will bind your app in any interface available, localhost does not. There is one post that describes the difference between localhost and 0.0.0.0, if I find I will update this post.