Add prediction endpoint container to k8s.yaml
Opened this issue · 0 comments
j-knudson commented
previous: #24
Background
We currently have a k8s.yaml script to run a series of containers in a pod. The prediction endpoint has been containerized and is available on docker hub. That container has been added to the VPS and is available to run in a development setting.
Technical Details
We have an Nginx container currently exposed on port 9000. The internal port for the prediction container is 5000. Configure the new container to be a part of the pod. Create a connection using WSGI (such as gunicorn) and running Nginx as a reverse proxy.
Potentially helpful links:
Flask Documentation
Acceptance Criteria
- A prediction endpoint container is a part of the pod
- The container can be used to generate a prediction
- It is configured for a production deployment
- The wiki has been updated with details on using the new container