When using the Databricks SDK to retrieve metadataโsuch as catalogs, schemas, or tablesโthrough its built-in API endpoints, does this incur any cost similar to running SQL queries?
Specifically, executing SQL queries via the API spins up a compute cluster (such as a SQL warehouse), which results in usage-based charges. The question is whether accessing metadata through the SDK's native endpoints also triggers compute resources in the same way, or if it avoids those costs entirely. How the native endpoints actually work?