Skip to content


deployment_name: Optional[str]=None,
make_production_ready: bool=False,
ignore_guardrails: bool=False,
use_case_name: Optional[str]=None
) -> Deployment


Create a deployment of a feature list. With a deployment, you can serve the feature list in production by either online or batch serving.

This triggers the orchestration of the feature materialization into the online feature store. A feature list is deployed without creating separate pipelines or using different tools.

Deployment can be disabled at any time if the serving of the feature list is not needed anymore. Unlike the log and wait approach adopted by some feature stores, disabling the deployment of a feature list doesn’t affect the serving of its historical requests.


  • deployment_name: Optional[str]
    Name of the deployment, if not provided, the name will be generated automatically.

  • make_production_ready: bool
    default: False
    Whether to convert the feature to production ready if it is not production ready.

  • ignore_guardrails: bool
    default: False
    Whether to ignore guardrails when trying to promote features in the list to production ready status.

  • use_case_name: Optional[str]
    Name of the use case associated with the deployment.


  • Deployment
    Deployment object of the feature list. The created deployment is disabled by default.


>>> feature_list = catalog.get_feature_list("invoice_feature_list")
>>> deployment = feature_list.deploy(
...   deployment_name="new deploy",
...   make_production_ready=True,
...   use_case_name="fraud detection",
... )

See Also