Skip to content

TestInitialSetup.md

Latest commit

 

History

History
121 lines (73 loc) · 5.52 KB

TestInitialSetup.md

File metadata and controls

121 lines (73 loc) · 5.52 KB

Testing the initial setup

Step 1. In the main branch, supply an explicit value or accept the defaults in the file, config/config.yaml. The pipelines uses multiple variables and they should be set for both 'pr' and 'dev' plus any additional environments. Also, set the variables for all models (i.e. london_taxi, docker_taxi, sequence_model). The config.yaml file is split into the following sections, set the values in each section:

  • aml_config: Stores the configuration of azure resources hosting the Azure Machine Learning workspace.
  • environment_config: Stores the base image and dynamic properties set at runtime.
  • pipeline_configs: Stores the configuration for pr and dev pipelines for each model supported by the solution.
  • deploy_configs: Stores online and batch configuration for deployments for each model.

Azure DevOps Steps (legacy/deprecated)

Azure Pipelines are deprecated and triggers are disabled; use the GitHub Workflows Steps below for active CI/CD. The steps in this section are retained for historical reference only.

Step 1. Provision Infrastructure - Execute the infrastructure provisioning pipeline (infra_provision_bicep_pipeline.yml OR infra_provision_terraform_pipeline.yml).

  • .azure-pipelines/infra/bicep/infra_provision_bicep.yml
  • .azure-pipelines/infra/bicep/infra_provision_terraform.yml

Step 2. Register Data Assets - Execute the register data asset pipeline (register_data_assets.yml).

  • .azure-pipelines/register_data_assets.yml

Step 3. Run CI workflow for a model of your choice - Execute the GitHub Actions workflows listed below for build validation Pipeline Parameters:

  • exec_environment: The environment to run the workflow in. Set to "ci" by default
  • model_type: The type of model for which to run the CI workflow. Set to {model name} by default

Pipelines:

  • .azure-pipelines/london_taxi_ci_pipeline.yml
  • .azure-pipelines/docker_taxi_ci_pipeline.yml

Step 4. Run CD pipeline for a model of your choice - Execute any of the Azure Pipelines created above for continuous integration.

Pipeline Parameters:

  • exec_environment: The environment to run the workflow in. Set to "ci" by default
  • model_type: The type of model for which to run the CI workflow. Set to {model name} by default

Pipelines:

  • .azure-pipelines/london_taxi_cd_pipeline.yml
  • .azure-pipelines/docker_taxi_cd_pipeline.yml

Github Workflows Steps

Step 1. Provision Infrastructure - Execute the infrastructure provision pipeline

  • .github/workflows/ModelFactory-Bicep-Deployment.yml
  • .github/workflows/ModelFactory-Terraform-Deployment.yml

Step 2. Register Data Assets - Execute the register data asset pipeline (register_data_assets.yml).

  • .github/workflows/register_data_assets.yml

Step 3. Run CI pipeline for a model of your choice - Execute any of the Azure Pipelines created above for build validation

Pipeline Parameters:

  • exec_environment: The environment to run the workflow in. Set to "ci" by default
  • model_type: The type of model for which to run the CI workflow. Set to {model name} by default

Workflows:

  • .github/workflows/london_taxi_ci_pipeline.yml
  • .github/workflows/docker_taxi_ci_pipeline.yml

Step 4. Run CD workflow for a model of your choice - Execute the GitHub Actions workflows listed below for deployment validation. Pipeline Parameters:

  • exec_environment: The environment to run the workflow in. Set to "cd" by default
  • model_type: The type of model for which to run the CI workflow. Set to {model name} by default

Workflows:

  • .github/workflows/london_taxi_cd_pipeline.yml
  • .github/workflows/docker_taxi_cd_pipeline.yml

Below is the sample job run for the CI Pipeline.

sample_ci_job

Preparing to Extend the solution

Step 1. Cleanup unneeded directories and files.

If you used Azure DevOps and Bicep for the infrastructure provisioning and data registration, delete the following directories/files:

  • ./github
  • /infra/terraform
  • .azure-pipelines/terraform

If you used Azure DevOps and Terraform for the infrastructure provisioning, delete the following directories/files

  • ./github
  • /infra/bicep
  • .azure-pipelines/bicep

If you used Github Actions and Bicep for the infrastructure provisioning, delete the following directories/files

  • ./azure-pipelines
  • /infra/terraform
  • .github/infra_provision_terraform.yml

If you used Github Actions and Terraform for the infrastructure provisioning, delete the following directories/files

  • ./azure-pipelines
  • /infra/bicep
  • .github/infra_provision_bicep.yml

Your team already has a process for provisioning data needed by your models, therefore, delete the data configuration files included in the solution for testing purposes

  • config/data_config.json

Step 2. Since you can confidently build the infrastructure, register data assets, and run a ci and cd build for a model, you are ready to add your own use cases to the Model Factory. See Onboarding a New Model

Known Issues

  • Currently, there are two file formats in use for configuration files. We have logged a bug against this and will make the format consistent in an upcoming release. See Different Config file formats
  • Use of a single configuration file could become a usability issue when there are many models supported by the solution, accordingly, we have raised an enhancement request to modularize the model configuration. See Many Models Configuration Usability Issue.