Feature Service Multiple Feature Views, each from different data source types (Spark, Snowflake, Postgres)

[oonisim]

Regarding Feature Service:

A feature service is an object that represents a logical group of features from one or more feature views.

[Inquiry]
Please confirm a Feature Service in Feast can group multiple feature views that come from completely different data sources, including Spark, Snowflake, Teradata, Postgres, and others.

There are Data Sources not directly implemented but contributed such as Terradata. Please confirm such external contribution Data Source will be also supported by the Feature Service.

The example code in Feature Service only use Sqlite local datasource only. Hence, it is not clear the scope of Feature Store support.

[oonisim]

Currently there is only one offline_store can be defined in feature_store.yaml. So only one Data Store per project should be the limitation.

Kindly update the documentation to:

A feature service is an object that represents a logical group of features from one or more feature views but only from single data source as only one offline_store can be specified in a FEAST project.