How-to Guides

How-to guides are task-oriented recipes for users who already have the plugin installed and configured. Each guide solves a specific problem.

Core workflows

• Schedule Pipelines: Configure cron and recurrence schedules for recurring Azure ML jobs.
• Use Data Assets: Integrate Azure ML Data Assets into your Kedro catalog with AzureMLAssetDataset and AzureMLPipelineDataset.
• Compile and Inspect: Generate Azure ML Pipeline YAML definitions and inspect them before submitting.

Advanced features

• Run Distributed Training: Scale Kedro nodes across multiple GPU instances with @distributed_job.
• Use MLflow: Track experiments with kedro-mlflow during Azure ML pipeline runs.
• Configure Multiple Workspaces: Target dev, staging, and production workspaces from a single configuration.
• Build a Custom Environment: Create and register an Azure ML environment with your project's dependencies.
• Deploy from CI/CD: Submit pipeline jobs from GitHub Action or other CI/CD systems.

Operations

• Authenticate: Configure Azure credentials for local development, CI/CD, and Azure ML compute.
• Troubleshoot: Diagnose common errors and debug failed pipeline runs.
• Contribute: Set up a development environment and contribute to the project.