Skip to main content

Outer Loop (Coming soon)

warning

Coming soon

The outer loop in Managed Azure ML takes over where the inner loop ends.
It focuses on promotion, validation, and production deployment of models that have already been captured and tagged in Azure ML during the inner loop stage.

The outer loop consists of two main GitHub workflows:

1. Outer Loop Test​

The outer loop test workflow promotes a model from the development workspace to the test workspace.

  • Purpose: Allows additional validation, performance checks, and QA in an environment that mirrors production but is isolated from live users.
  • Default trigger: Runs automatically on merge to the main branch. This ensures that all mainline code is validated in a test environment before production promotion.
  • Availability: Only used in the full end-to-end architecture tier.

outer loop test Outer loop test

warning

Link to outer loop test workflow file needs to added once ready.

2. Outer Loop Prod​

The outer loop prod workflow promotes a model into the production workspace.

How it works depends on your chosen architecture tier:

  • Simple end-to-end: Promotes directly from development to production.

  • Full end-to-end: Promotes from test to production, after the outer loop test workflow completes successfully.

  • Default trigger: Manual release from GitHub. This gives you full control over when a model is deployed to production.

outer loop test Outer loop prod

warning

Link to outer loop prod workflow file needs to added once ready.