Skip to content

11. Deploy and Serve

In this tutorial, you'll learn how to:

Step 1: Mark a Feature as Production-Ready

  1. Navigate to the Features Catalog under the Experiment section of the menu. Name

  2. Locate the feature "NEW_APPLICATION_EXT_SOURCE_2". Click the three dots next to its current status (e.g., DRAFT) Readiness Tree Dots. Change its readiness to Upgrade Icon. Name

  3. Confirm the change. Name

Note

If your catalog has an Approval Flow enabled, promoting a feature to "Production-Ready" requires approval.

Step 2: Enable the Approval Flow

  1. Click the settings icon Settings icon next to the catalog name at the top of the screen. Name

  2. In catalog settings, enable the 'Approval Flow' option and save changes. Name

  3. A validation mark next to the catalog name indicates that Approval Flow is active. Name

Step 3: Request Production-Ready Upgrade

  1. Search for the feature "CLIENT_Max_of_BureauReportedCredits_AMT_CREDIT_SUM_DEBT_To_AMT_CREDIT_SUMs_104w" in the Feature Catlog. Click Upgrade Icon and confirm the readiness change. Name

  2. This action triggers an approval request. Click View Button to open the request. Name

  3. You’ll be redirected to Approve → Feature Requests → New Production-Ready Feature. Name

Step 4: Review Checks

Click the request to view all automated checks.

Note

The system runs several checks — such as compliance with default cleaning operations, validation of feature job setting, table status and backtesting to prevent training-serving inconsistencies.

Name

If any checks fail (e.g., missing backtests):

  1. Click Run Analysis Button to run a "Data Availability and Freshness Analysis".
  2. Then click Run Backtest Button to run a backtest.

Once complete, all checks should turn green.

Name

Step 5: Review Feature

Review the feature definition and its use of tiles (partial aggregations).

Name

Name

Step 6: Request Review

When ready, click Request Review Button to submit for review.

Your request now awaits approval.

Name

Step 7: Approve and Merge Request

Click Approve Request Button to approve.

Name

Click Merge Button to merge.

Step 8: Audit Merged Request

View merged requests under the Merged tab.

Name

Step 9: Requests by batch

To upgrade all features in a feature list to Production-Ready.

  1. Go to the Feature List Catalog.

  2. Click Check Button next to Calculator Button of the feature list you want to upgrade. Name

  3. Confirm the action. Name

  4. In the pop-up, click View Button. Name

  5. You’ll be redirected to Approve → Feature Requests → New Production-Ready Feature. Name

  6. Run further backtests if needed. Name

  7. Once all checks passed, select all requests on each page, and then request reviews in batches. Name

  8. Approve and merge requests. Name

    Name

Merged requests appear under the Merged tab.

Name

Step 10: Deploy and Serve Feature List

Once all features are Production-Ready, proceed to deploy the list.

  1. Initiate Deployment: Click Rock Icon to initiate deployment. Name

  2. Configure Deployment Details: Provide a descriptive name (e.g. New Application: 308 Features from XGBoost [293 features: Loan Default Risk Assessment] by cumulative Feature Key Importance (0.95)) and associate the deployment with your Use Case for tracking. Name

  3. View the Deployment: Click View Deployments to access the deployment in the Deployments Catalog under the Manage section of the menu.

  4. Enable the Deployment:

    Locate the new deployment and select it. Name

    Click the three dots next to the 'NEW' label and choose 'Enable.' Name Name

  5. Compute a Batch Feature Table: In the Batch Feature Tables tab, compute a batch feature table from a source table or a managed view that contains the entity ids of your use case. Name

Note

Online store functionality is not available in the Hosted Tutorials.

If you configured an online store, follow these steps to enable online serving:

  1. Click Settings icon next to the Catalog name at the top of the screen. Name

  2. In the Online Serving tab, you’ll find ready-to-use Python and shell script templates for deploying REST API services.

    Name

    Name

Step 11: Deploy and Serve Model

Now deploy the model associated with your feature list.

  1. Navigate to the Models Catalog and select the model.

    Name

  2. Open the About tab, click Deploy Model button. Provide a name (e.g. New Application: XGBoost with top 308 keys)

    Name

  3. Navigate to the Deployments Catalog, locate your model and enable it.

    Name

  4. Click the model deployment to view details.

    Name