You are developing a Snowflake Native Application that relies on external functions (UDFs/UDTFs) for data processing. A new version of your application includes significant changes to these external functions, potentially affecting existing consumer data and queries. How should you manage the upgrade process to minimize disruption and ensure data integrity in consumer accounts? Assume consumers rely on your function names in their queries. Which of the following approach is MOST suitable?

• A. Use release directives to direct users to an intermediate version that provides a compatibility layer for the external function changes. This layer translates calls to the old function names into calls to the new functions, allowing for a phased transition.
• B. Rename all the external functions in the new version and instruct consumers to update their queries accordingly after the upgrade.
• C. Update the external functions in place with the new logic. Snowflake's versioning will automatically handle any compatibility issues for existing queries.
• D. Create new versions of the external functions with different names while maintaining the old functions in the new version. Consumers can then migrate to the new functions at their own pace. Use release directives to route new users to new functions.
• E. Develop wrapper functions in the new version that call the original external functions with the new logic applied. Consumers can then gradually migrate to the wrapper functions without immediately changing their queries.

You are developing a Snowflake Native Application that utilizes Snowpark Container Services (SPCS). Your application requires a Compute Pool configured with autoscaling capabilities to handle fluctuating workloads. However, you want to minimize costs when the application is idle. Which of the following considerations are MOST important when defining the 'MIN NODES' and 'MAX NODES' parameters for your Compute Pool, along with the 'AUTO RESUME setting?

• A. Setting to 0 allows the Compute Pool to scale down completely when idle, eliminating costs. 'AUTO_RESUME should be set to TRUE to automatically start the pool when a new request is received.
• B. The 'AUTO_RESUME parameter is irrelevant as the compute pool automatically scales based on current workload. should be set to the expected average workload.
• C. The optimal configuration depends on the balance between cost and performance. A higher 'MIN_NODES' value provides faster initial response times but incurs higher costs when idle. 'AUTO RESUME should be set to TRUE if you want the pool to automatically restart based on incoming requests.
• D. Compute Pool autoscaling is managed entirely by the underlying container orchestration system and Snowflake only provides basic resource allocation. The 'MIN NODES' and 'MAX NODES parameters are advisory only.
• E. Setting to a non-zero value ensures that some resources are always available, leading to faster response times for initial requests. 'AUTO RESUME' should be set to FALSE to prevent unnecessary resource usage.

A Snowflake Native Application developer, Bob, is using Snowsight to manage access to the application. He has created a custom role, , and wants to grant this role the ability to grant and revoke privileges on all objects within the application's schema, app_data'. Which of the following Snowsight actions (or sequence of actions) is the MOST effective and secure way to achieve this?
(Assume Bob is logged in as a user with 'ACCOUNTADMIN' role).

• A. Within Snowsight, navigate to the 'app_data' schema, select 'Grant Privileges', grant the 'USAGE privilege to the and then individually grant 'SELECT, 'INSERT, 'UPDATE' , and "DELETE privileges on each table and view within the schema.
• B. Within Snowsight, navigate to the application database, select 'Grant Privileges', and grant the 'USAGE' privilege to the .
• C. Within Snowsight, navigate to the 'app_data' schema, select 'Grant Privileges', and grant the 'OWNERSHIP' privilege to the .
• D. Grant the the 'APPLY MASKING POLICY privilege globally at the account level using Snowsight.
• E. In Snowsight, write and execute the following SQL command:

You are managing the release of a new version (v2.0) of your Snowflake Native Application that introduces significant schema changes to shared tables accessed by consumer accounts. To minimize disruption during the upgrade, you want to implement a phased rollout while ensuring backward compatibility for consumers who haven't yet upgraded. Which of the following strategies would allow you to achieve this goal effectively? (Select TWO)

• A. Use dynamic SQL within stored procedures to conditionally access either the old or new schema based on the consumer's account ID, identified through a custom configuration table.
• B. Create separate versions of the shared tables for vl .0 and v2.0, using views to present a unified interface. Grant consumer 'USAGE on view after upgrading.
• C. Create views on top of the new schema in v2.0 that emulate the old schema from vl .0. Grant consumers 'USAGE' on these views until they upgrade.
• D. Implement API versioning within your application code and provide compatibility layers to handle requests from consumers running older versions.
• E. Publish v2.0 without any schema changes and rely on consumers to manually migrate their data to new tables after upgrading.

A provider is developing a Snowflake Native Application that shares event data with consumers. The provider wants to ensure that specific columns in the event data are obfuscated for certain consumer accounts, while remaining visible to others. Which of the following approaches is the MOST secure and maintainable way to achieve this?

• A. Implement dynamic data masking policies on the sensitive columns using the CURRENT_ACCOUNT() function to determine which accounts receive obfuscated data.
• B. Use a stored procedure to query the event data, applying obfuscation logic based on the consumer's APP INSTANCE ID() before returning the results.
• C. Implement conditional logic within the event table function to obfuscate data based on the consumer's APP_INSTANCE_ID().
• D. Create a separate view for each consumer account with the sensitive columns masked using a CASE statement based on the CURRENT_ACCOUNT() function.
• E. Data masking is not supported in Native Apps. The provider must rely on consumer-side masking implementations.