Question

A GenAI engineer has registered version 3 of a RAG chain in Unity Catalog under the name main.genai.support_bot. They now need to serve this specific model version behind a Databricks Model Serving endpoint using the Python SDK, ensuring the endpoint routes 100% of traffic to that version. Which approach correctly deploys the model?

Accepted Answer

Create a serving endpoint with a served_entities entry specifying entity_name='main.genai.support_bot', entity_version='3', and a traffic_config route with traffic_percentage=100.

Answer

Log the chain again with mlflow.pyfunc.log_model() inside the endpoint creation call so the serving runtime can build the container from source.

Answer

Register the model to the Databricks Feature Store and reference the feature table name as the served entity in the endpoint config.

Answer

Set the endpoint's environment variable MODEL_VERSION=3 and point entity_name to the workspace model registry path without a version.