Snowflake's virtual warehouses are the core of its data analytics platform, allowing users to separate compute from storage and handle computationally-intensive tasks such as querying, loading, or manipulating data. Warehouses come in various sizes, ranging from small to extremely large, with larger warehouses costing more money than smaller ones. Resizing a warehouse can improve performance by providing more compute resources for operations like complex queries, data loading, and unloading. However, this requires manual sizing, which can be time-consuming and expensive. To handle concurrency, users have two options: make their warehouse larger or set up multiple warehouses running in parallel. While there is no limit to the number of virtual warehouses that can be created, managing multiple warehouses can become expensive quickly. Multi-cluster warehouses offer advantages like improved concurrency but come with drawbacks such as limited direct access to tweak warehouse size and potential costs. Propel Data Cloud's analytics backend system provides an alternative solution for building in-product analytics using Snowflake as a data warehouse, offering easy-to-use GraphQL APIs without the need for manual multi-cluster warehouse setup.