After successfully deploying a machine learning model in a Databricks environment using MLflow, the model is now handling a high volume of requests. To ensure the model's high availability, what strategy should the data scientist adopt?