Machine learning preview feature

Question

Hi, curious about what is a preview feature in azure? I want to use it in my prod environment but I have concern that if there is any concern with preview feature . Thank you for any insight

Answer 1

Hi,

Via

https://learn.microsoft.com/en-us/azure/machine-learning/how-to-enable-preview-features?view=azureml-api-2

In Azure Machine Learning, new features and improvements are often first released as preview features before they're made generally available (GA). As new features are introduced, you can turn them on or off in the Azure Machine Learning studio at your convenience. That way, you get a chance to use the latest features, evaluate how they fit your work needs and provide feedback to shape the product. Your feedback is very valuable and it helps us constantly improve the product.

Some preview features provide access to entire new functionality while others may reflect a change to the user interface, but little or no change in functionality.

I hope this helps.

Answer 2

Hello @Roshan Murthy

Thanks for reaching out to us.

Preview features in Azure represent new capabilities or services that Microsoft offers to users before they are officially released as general availability (GA). These features are intended to allow users to test and provide feedback on upcoming functionality while Microsoft continues to refine and improve them. Here’s a detailed look at what preview features are, their potential benefits, and the concerns associated with using them:

What is a Preview Feature?

• Early Access: Preview features give users early access to new functionalities that are still under development. They are usually available for testing and feedback before the feature is fully released.
• Beta Stage: They are often in a beta or early release stage, meaning they might still be undergoing testing and may have bugs or incomplete functionality.

Benefits of Using Preview Features

1. Early Adoption: You get to leverage new functionalities and capabilities before they are widely available.
2. Feedback Opportunity: You can provide feedback to Microsoft, potentially influencing the final design and functionality of the feature.
3. Competitive Edge: Early access to advanced features can give you a competitive advantage in your market.

Concerns and Considerations

1. Stability and Reliability:
   • Risk of Bugs: Preview features may have unresolved bugs or stability issues. They might not be as reliable as GA features.
   • Frequent Changes: The feature might undergo significant changes based on user feedback, which can impact your production environment.
2. Support and Documentation:
   • Limited Support: Microsoft might offer limited support for preview features. Full support is generally provided only once the feature is in GA.
   • Documentation: Documentation might be incomplete or subject to change as the feature evolves.
3. Performance:
   • Inconsistent Performance: Performance might vary and may not be as optimized as GA features. There might be performance issues that could affect your production workloads.
4. Integration:
   • Compatibility: Preview features may not be fully compatible with other Azure services or third-party tools. Integration issues can arise, which could impact your overall architecture.
5. Feature Deletion:
   • Deprecation: Preview features can be deprecated or removed before they reach GA. There’s a risk that the feature might not be available in the future or that you may need to make changes to adapt to the GA version.
6. Compliance and Security:
   • Compliance: Ensure that the preview feature complies with your organization’s regulatory and security requirements. There may be additional risks associated with using features that are not fully vetted.

Best Practices for Using Preview Features

1. Testing Environment:
   • Use Non-Production Environments: Test preview features in a staging or development environment before considering their use in production. This helps identify any issues without impacting your live systems.
2. Backup and Contingency Plans:
   • Have Rollback Plans: Ensure you have a rollback plan or contingency strategy if the preview feature causes issues in your environment.
3. Stay Informed:
   • Follow Updates: Keep track of updates and changes to the preview feature through Azure updates, documentation, and release notes.
4. Provide Feedback:
   • Engage with Microsoft: Provide feedback and report issues to help improve the feature and ensure it meets your needs when it reaches GA.
5. Review Terms and Conditions:
   • Understand Limitations: Review any terms and conditions associated with using preview features to understand potential limitations and obligations.

Example Scenario

If you're considering using a preview machine learning feature, such as a new model type or an advanced training capability:

1. Test Thoroughly: Evaluate the feature's performance and compatibility with your existing workflows.
2. Monitor Impact: Keep an eye on how the feature affects your system's performance and stability.
3. Plan for Changes: Be prepared for changes or updates in the feature as it evolves toward GA.

I hope this helps.

Regards, Yutong -Please kindly accept the answer if you feel helpful to support the community, thanks a lot.