Skip to content
This repository has been archived by the owner on Jul 22, 2024. It is now read-only.

Continuous Retrain and Deployment in Production #16

Open
meethariprasad opened this issue Apr 1, 2020 · 0 comments
Open

Continuous Retrain and Deployment in Production #16

meethariprasad opened this issue Apr 1, 2020 · 0 comments

Comments

@meethariprasad
Copy link

meethariprasad commented Apr 1, 2020

Scenario Description

  • There is one model that need to be daily trained and updated through scheduled pipelines.

  • The new version of the model will also be updated in existing ACI REST API in staging instance.

  • Please note we don't want REST API to be created again, as there are applications using existing URL.

  • Also, if update failed, REST API to retain its state, not error out when called.

  • How to move this entire scenario to production deployment, so that in production once "deployed", pipelines can run seamlessly and update model and update existing production REST API with new model version?

  • Please note we don't want REST API to be created again, as there are applications using existing URL. Also, if update failed, REST API to retain its state, not error out when called.

Reasons for scenario

This is a extremely important business requirement and highly common scenario. There is no clear document|tutorial|guidelines around to do this.

Other Comments

Sign up for free to subscribe to this conversation on GitHub. Already have an account? Sign in.
Labels
None yet
Projects
None yet
Development

No branches or pull requests

1 participant