DP-700 Practice Test

Answer choices

1. A. Create a new, dedicated Spark pool with a larger number of driver nodes and a higher memory allocation per executor, and then assign this pool to all existing data pipelines.
2. B. Increase the auto-scaling limits of the default Spark pool to dynamically adjust the number of executors based on workload demand, while maintaining the existing memory allocation per executor.
3. C. Configure a custom Spark pool with a fixed number of executors and a higher memory allocation per executor, based on the observed peak resource utilization during initial testing, and assign this pool to the data pipeline.
4. D. Implement a cost-saving strategy by reducing the number of executors in the default Spark pool and monitoring performance; if issues arise, increase the number of executors as needed.

Wrong-answer review

• A. Create a new, dedicated Spark pool with a larger number of driver nodes and a higher memory allocation per executor, and then assign this pool to all existing data pipelines.: Create a new, dedicated Spark pool with a larger number of driver nodes and a higher memory allocation per executor, and then assign this pool to all existing data pipelines. is incorrect here because it conflicts with the supported Fabric approach for this scenario. The most effective strategy is to create a custom Spark pool with a fixed number of executors and increased memory per executor. Streaming workloads often benefit from predictable resource allocation to avoid unpredictable performance fluctuations. Although auto-scaling can be useful, it introduces latency as the cluster scales up and down, which can negatively impact streaming data processing. A fixed, appropriately sized pool provides consistent performance and avoids the overhead of dynamic scaling. The team's initial testing provides valuable data to determine the optimal number of executors and memory allocation. Assigning this custom pool to the specific data pipeline ensures it receives the necessary resources without impacting other workloads in the workspace.
• B. Increase the auto-scaling limits of the default Spark pool to dynamically adjust the number of executors based on workload demand, while maintaining the existing memory allocation per executor.: Increase the auto-scaling limits of the default Spark pool to dynamically adjust the number of executors based on workload demand, while maintaining the existing memory allocation per executor. is incorrect here because it conflicts with the supported Fabric approach for this scenario. The most effective strategy is to create a custom Spark pool with a fixed number of executors and increased memory per executor. Streaming workloads often benefit from predictable resource allocation to avoid unpredictable performance fluctuations. Although auto-scaling can be useful, it introduces latency as the cluster scales up and down, which can negatively impact streaming data processing. A fixed, appropriately sized pool provides consistent performance and avoids the overhead of dynamic scaling. The team's initial testing provides valuable data to determine the optimal number of executors and memory allocation. Assigning this custom pool to the specific data pipeline ensures it receives the necessary resources without impacting other workloads in the workspace.
• D. Implement a cost-saving strategy by reducing the number of executors in the default Spark pool and monitoring performance; if issues arise, increase the number of executors as needed.: Implement a cost-saving strategy by reducing the number of executors in the default Spark pool and monitoring performance; if issues arise, increase the number of executors as needed. is incorrect here because it conflicts with the supported Fabric approach for this scenario. The most effective strategy is to create a custom Spark pool with a fixed number of executors and increased memory per executor. Streaming workloads often benefit from predictable resource allocation to avoid unpredictable performance fluctuations. Although auto-scaling can be useful, it introduces latency as the cluster scales up and down, which can negatively impact streaming data processing. A fixed, appropriately sized pool provides consistent performance and avoids the overhead of dynamic scaling. The team's initial testing provides valuable data to determine the optimal number of executors and memory allocation. Assigning this custom pool to the specific data pipeline ensures it receives the necessary resources without impacting other workloads in the workspace.

Answer choices

1. A. Configure Row-Level Security (RLS) using T-SQL functions to filter the 'CustomerSSN' column based on user roles.
2. B. Implement Dynamic Data Masking (DDM) rules on the 'CustomerSSN' column to mask the data based on user roles and data classifications.
3. C. Create a data access role with read permissions only on a masked view of the 'Transactions' table, excluding direct access to the base table.
4. D. Apply a sensitivity label to the 'Transactions' table and configure data loss prevention (DLP) policies to prevent the export of the 'CustomerSSN' column.

Wrong-answer review

• A. Configure Row-Level Security (RLS) using T-SQL functions to filter the 'CustomerSSN' column based on user roles.: Configure Row-Level Security (RLS) using T-SQL functions to filter the 'CustomerSSN' column based on user roles. is incorrect here because it conflicts with the supported Fabric approach for this scenario. Dynamic Data Masking (DDM) is the most appropriate solution for this scenario. DDM allows you to define rules that mask sensitive data based on user roles or data classifications without modifying the underlying data. This enables analysts to work with masked versions of the data while protecting the raw SSN values from unauthorized access. The configuration is directly within the Fabric Warehouse, aligning with the requirement. RLS would filter rows, not mask columns. Creating a masked view is a viable option, but DDM is more flexible and centrally managed. DLP policies are for preventing data exfiltration, not for controlling access within the Fabric environment.
• C. Create a data access role with read permissions only on a masked view of the 'Transactions' table, excluding direct access to the base table.: Create a data access role with read permissions only on a masked view of the 'Transactions' table, excluding direct access to the base table. is incorrect here because it conflicts with the supported Fabric approach for this scenario. Dynamic Data Masking (DDM) is the most appropriate solution for this scenario. DDM allows you to define rules that mask sensitive data based on user roles or data classifications without modifying the underlying data. This enables analysts to work with masked versions of the data while protecting the raw SSN values from unauthorized access. The configuration is directly within the Fabric Warehouse, aligning with the requirement. RLS would filter rows, not mask columns. Creating a masked view is a viable option, but DDM is more flexible and centrally managed. DLP policies are for preventing data exfiltration, not for controlling access within the Fabric environment.
• D. Apply a sensitivity label to the 'Transactions' table and configure data loss prevention (DLP) policies to prevent the export of the 'CustomerSSN' column.: Apply a sensitivity label to the 'Transactions' table and configure data loss prevention (DLP) policies to prevent the export of the 'CustomerSSN' column. is incorrect here because it conflicts with the supported Fabric approach for this scenario. Dynamic Data Masking (DDM) is the most appropriate solution for this scenario. DDM allows you to define rules that mask sensitive data based on user roles or data classifications without modifying the underlying data. This enables analysts to work with masked versions of the data while protecting the raw SSN values from unauthorized access. The configuration is directly within the Fabric Warehouse, aligning with the requirement. RLS would filter rows, not mask columns. Creating a masked view is a viable option, but DDM is more flexible and centrally managed. DLP policies are for preventing data exfiltration, not for controlling access within the Fabric environment.

Answer choices

1. A. Upgrade the workspace to a 'Trial' license to gain access to more compute resources and monitor performance.
2. B. Upgrade the workspace to a 'Premium' license and allocate a Fabric capacity node (F SKU) to provide dedicated compute resources.
3. C. Configure the workspace to use a default Spark pool with auto-scaling enabled to dynamically adjust compute resources based on workload.
4. D. Migrate the data loading and transformation processes to a different workspace with a 'Standard' license to distribute the workload.

Wrong-answer review

• A. Upgrade the workspace to a 'Trial' license to gain access to more compute resources and monitor performance.: Upgrade the workspace to a 'Trial' license to gain access to more compute resources and monitor performance. is incorrect here because it conflicts with the supported Fabric approach for this scenario. Upgrading the workspace to a 'Premium' license and allocating an F SKU is the correct solution. The 'Free' license severely limits compute resources, leading to timeouts. 'Trial' licenses also have limitations. 'Premium' licenses unlock the ability to allocate dedicated F SKUs, providing the necessary compute power for demanding workloads. Although auto-scaling can be beneficial, it's not a guaranteed solution for immediate performance issues and requires careful configuration. Migrating to another workspace introduces unnecessary complexity and doesn't address the root cause of the resource constraints.
• C. Configure the workspace to use a default Spark pool with auto-scaling enabled to dynamically adjust compute resources based on workload.: Configure the workspace to use a default Spark pool with auto-scaling enabled to dynamically adjust compute resources based on workload. is incorrect here because it conflicts with the supported Fabric approach for this scenario. Upgrading the workspace to a 'Premium' license and allocating an F SKU is the correct solution. The 'Free' license severely limits compute resources, leading to timeouts. 'Trial' licenses also have limitations. 'Premium' licenses unlock the ability to allocate dedicated F SKUs, providing the necessary compute power for demanding workloads. Although auto-scaling can be beneficial, it's not a guaranteed solution for immediate performance issues and requires careful configuration. Migrating to another workspace introduces unnecessary complexity and doesn't address the root cause of the resource constraints.
• D. Migrate the data loading and transformation processes to a different workspace with a 'Standard' license to distribute the workload.: Migrate the data loading and transformation processes to a different workspace with a 'Standard' license to distribute the workload. is incorrect here because it conflicts with the supported Fabric approach for this scenario. Upgrading the workspace to a 'Premium' license and allocating an F SKU is the correct solution. The 'Free' license severely limits compute resources, leading to timeouts. 'Trial' licenses also have limitations. 'Premium' licenses unlock the ability to allocate dedicated F SKUs, providing the necessary compute power for demanding workloads. Although auto-scaling can be beneficial, it's not a guaranteed solution for immediate performance issues and requires careful configuration. Migrating to another workspace introduces unnecessary complexity and doesn't address the root cause of the resource constraints.

Answer choices

1. A. Grant the 'Sales Insights' workspace 'Contributor' access to the 'Data Refinery' workspace, allowing them to directly query the lakehouse.
2. B. Create an internal OneLake shortcut in the 'Sales Insights' workspace that points to the specific curated lakehouse folder within the 'Data Refinery' workspace.
3. C. Configure a data access role in the 'Data Refinery' workspace and assign it to users in the 'Sales Insights' workspace, granting them direct access to the lakehouse.
4. D. Copy the data from the 'Data Refinery' workspace to a new folder within the 'Sales Insights' workspace, creating a duplicate dataset.

Wrong-answer review

• A. Grant the 'Sales Insights' workspace 'Contributor' access to the 'Data Refinery' workspace, allowing them to directly query the lakehouse.: Grant the 'Sales Insights' workspace 'Contributor' access to the 'Data Refinery' workspace, allowing them to directly query the lakehouse. is incorrect here because it conflicts with the supported Fabric approach for this scenario. Creating an internal OneLake shortcut is the correct approach. Shortcuts allow you to access data residing in another workspace without granting direct access or duplicating data. This maintains the security boundaries of the 'Data Refinery' workspace while providing the 'Sales Insights' workspace with a secure, logical view of the required data. Shortcuts inherit the access controls of the source data, ensuring that only authorized users can access the data through the shortcut. This aligns with the principle of least privilege and simplifies data governance.
• C. Configure a data access role in the 'Data Refinery' workspace and assign it to users in the 'Sales Insights' workspace, granting them direct access to the lakehouse.: Configure a data access role in the 'Data Refinery' workspace and assign it to users in the 'Sales Insights' workspace, granting them direct access to the lakehouse. is incorrect here because it conflicts with the supported Fabric approach for this scenario. Creating an internal OneLake shortcut is the correct approach. Shortcuts allow you to access data residing in another workspace without granting direct access or duplicating data. This maintains the security boundaries of the 'Data Refinery' workspace while providing the 'Sales Insights' workspace with a secure, logical view of the required data. Shortcuts inherit the access controls of the source data, ensuring that only authorized users can access the data through the shortcut. This aligns with the principle of least privilege and simplifies data governance.
• D. Copy the data from the 'Data Refinery' workspace to a new folder within the 'Sales Insights' workspace, creating a duplicate dataset.: Copy the data from the 'Data Refinery' workspace to a new folder within the 'Sales Insights' workspace, creating a duplicate dataset. is incorrect here because it conflicts with the supported Fabric approach for this scenario. Creating an internal OneLake shortcut is the correct approach. Shortcuts allow you to access data residing in another workspace without granting direct access or duplicating data. This maintains the security boundaries of the 'Data Refinery' workspace while providing the 'Sales Insights' workspace with a secure, logical view of the required data. Shortcuts inherit the access controls of the source data, ensuring that only authorized users can access the data through the shortcut. This aligns with the principle of least privilege and simplifies data governance.

Answer choices

1. A. Install and run the fabric-cicd npm package within a pipeline task to deploy the workspace items.
2. B. Use the native Azure DevOps Release Pipelines task for Power BI Service deployment.
3. C. Create a pipeline that copies the JSON definition files directly into the OneLake system folders via the Azure CLI.
4. D. Enable the 'Auto-Deploy to Workspace' option in the Azure DevOps repository settings.

Wrong-answer review

• B. Use the native Azure DevOps Release Pipelines task for Power BI Service deployment.: Use the native Azure DevOps Release Pipelines task for Power BI Service deployment. is incorrect here because it conflicts with the supported Fabric approach for this scenario. For automated CI/CD using Azure DevOps, you can run a script using the fabric-cicd tool (available as an npm package) to deploy workspace items from a Git repository to a Fabric workspace. The fabric-cicd tool compiles and deploys the Fabric item definitions. Power BI service deployment tasks do not support Fabric-native items like lakehouses. You cannot copy JSON definition files directly into OneLake system folders via Azure CLI to create items. There is no 'Auto-Deploy to Workspace' option in Azure DevOps repository settings for Fabric.
• C. Create a pipeline that copies the JSON definition files directly into the OneLake system folders via the Azure CLI.: Create a pipeline that copies the JSON definition files directly into the OneLake system folders via the Azure CLI. is incorrect here because it conflicts with the supported Fabric approach for this scenario. For automated CI/CD using Azure DevOps, you can run a script using the fabric-cicd tool (available as an npm package) to deploy workspace items from a Git repository to a Fabric workspace. The fabric-cicd tool compiles and deploys the Fabric item definitions. Power BI service deployment tasks do not support Fabric-native items like lakehouses. You cannot copy JSON definition files directly into OneLake system folders via Azure CLI to create items. There is no 'Auto-Deploy to Workspace' option in Azure DevOps repository settings for Fabric.
• D. Enable the 'Auto-Deploy to Workspace' option in the Azure DevOps repository settings.: Enable the 'Auto-Deploy to Workspace' option in the Azure DevOps repository settings. is incorrect here because it conflicts with the supported Fabric approach for this scenario. For automated CI/CD using Azure DevOps, you can run a script using the fabric-cicd tool (available as an npm package) to deploy workspace items from a Git repository to a Fabric workspace. The fabric-cicd tool compiles and deploys the Fabric item definitions. Power BI service deployment tasks do not support Fabric-native items like lakehouses. You cannot copy JSON definition files directly into OneLake system folders via Azure CLI to create items. There is no 'Auto-Deploy to Workspace' option in Azure DevOps repository settings for Fabric.

Answer choices

1. A. Grant the 'Reader' role at the OneLake root level to the marketing team and the 'Contributor' role to the finance team.
2. B. Create a custom Spark pool with specific permissions and assign it to the workspace, then configure access to the 'Sales Data' folder through the Spark pool’s settings.
3. C. Create a data access role named 'Marketing Analysts' with the 'Reader' permission and assign it to the marketing team. Create a data access role named 'Finance Team' with the 'Contributor' permission and assign it to the finance team, both scoped to the 'Sales Data' folder.
4. D. Modify the workspace's default data access settings to allow read-only access for all users and then grant the finance team explicit write permissions on the 'Sales Data' folder.

Wrong-answer review

• A. Grant the 'Reader' role at the OneLake root level to the marketing team and the 'Contributor' role to the finance team.: Grant the 'Reader' role at the OneLake root level to the marketing team and the 'Contributor' role to the finance team. is incorrect here because it conflicts with the supported Fabric approach for this scenario. The correct answer is to create data access roles scoped to the 'Sales Data' folder. Data access roles in OneLake (Preview) provide granular control over access permissions at the folder level. This allows the data engineering team to define specific roles ('Marketing Analysts' and 'Finance Team') with precisely the required permissions (read-only and contributor, respectively) and assign them to the appropriate teams. This approach minimizes the risk of unintended access and adheres to the principle of least privilege. Folder-level access control is a key feature for securing sensitive data within OneLake.
• B. Create a custom Spark pool with specific permissions and assign it to the workspace, then configure access to the 'Sales Data' folder through the Spark pool’s settings.: Create a custom Spark pool with specific permissions and assign it to the workspace, then configure access to the 'Sales Data' folder through the Spark pool’s settings. is incorrect here because it conflicts with the supported Fabric approach for this scenario. The correct answer is to create data access roles scoped to the 'Sales Data' folder. Data access roles in OneLake (Preview) provide granular control over access permissions at the folder level. This allows the data engineering team to define specific roles ('Marketing Analysts' and 'Finance Team') with precisely the required permissions (read-only and contributor, respectively) and assign them to the appropriate teams. This approach minimizes the risk of unintended access and adheres to the principle of least privilege. Folder-level access control is a key feature for securing sensitive data within OneLake.
• D. Modify the workspace's default data access settings to allow read-only access for all users and then grant the finance team explicit write permissions on the 'Sales Data' folder.: Modify the workspace's default data access settings to allow read-only access for all users and then grant the finance team explicit write permissions on the 'Sales Data' folder. is incorrect here because it conflicts with the supported Fabric approach for this scenario. The correct answer is to create data access roles scoped to the 'Sales Data' folder. Data access roles in OneLake (Preview) provide granular control over access permissions at the folder level. This allows the data engineering team to define specific roles ('Marketing Analysts' and 'Finance Team') with precisely the required permissions (read-only and contributor, respectively) and assign them to the appropriate teams. This approach minimizes the risk of unintended access and adheres to the principle of least privilege. Folder-level access control is a key feature for securing sensitive data within OneLake.

Answer choices

1. A. Share the 'SalesLake' item and grant only the 'Read' permission.
2. B. Share the 'SalesLake' item and grant the 'Read all OneLake data (ReadAll)' permission.
3. C. Assign the third-party service principal to the workspace 'Contributor' role.
4. D. Configure a OneLake data access role with 'ReadWrite' permissions at the folder level.

Wrong-answer review

• A. Share the 'SalesLake' item and grant only the 'Read' permission.: Share the 'SalesLake' item and grant only the 'Read' permission. is incorrect here because it conflicts with the supported Fabric approach for this scenario. The Read all OneLake data (ReadAll) permission allows external clients and compute engines, such as Azure Databricks, to read the underlying data in OneLake directly using ADLS Gen2 APIs. Granting only the 'Read' permission allows querying the SQL endpoint and the default semantic model, but does not grant direct file-level access. Assigning the 'Contributor' role provides workspace-level item management privileges, which violates the principle of least privilege. OneLake data access roles are used for granular folder-level permissions for workspace viewers, but ReadAll is the standard item sharing option for direct OneLake API access.
• C. Assign the third-party service principal to the workspace 'Contributor' role.: Assign the third-party service principal to the workspace 'Contributor' role. is incorrect here because it conflicts with the supported Fabric approach for this scenario. The Read all OneLake data (ReadAll) permission allows external clients and compute engines, such as Azure Databricks, to read the underlying data in OneLake directly using ADLS Gen2 APIs. Granting only the 'Read' permission allows querying the SQL endpoint and the default semantic model, but does not grant direct file-level access. Assigning the 'Contributor' role provides workspace-level item management privileges, which violates the principle of least privilege. OneLake data access roles are used for granular folder-level permissions for workspace viewers, but ReadAll is the standard item sharing option for direct OneLake API access.
• D. Configure a OneLake data access role with 'ReadWrite' permissions at the folder level.: Configure a OneLake data access role with 'ReadWrite' permissions at the folder level. is incorrect here because it conflicts with the supported Fabric approach for this scenario. The Read all OneLake data (ReadAll) permission allows external clients and compute engines, such as Azure Databricks, to read the underlying data in OneLake directly using ADLS Gen2 APIs. Granting only the 'Read' permission allows querying the SQL endpoint and the default semantic model, but does not grant direct file-level access. Assigning the 'Contributor' role provides workspace-level item management privileges, which violates the principle of least privilege. OneLake data access roles are used for granular folder-level permissions for workspace viewers, but ReadAll is the standard item sharing option for direct OneLake API access.

Answer choices

1. A. Configure the Git connection using a Personal Access Token (PAT) generated by the workspace administrator.
2. B. Authenticate the Git connection using a Service Principal registered in Microsoft Entra ID and configured in Azure DevOps.
3. C. Enable Tenant-level Single Sign-On (SSO) and delegate access via Azure DevOps organization-level access policies.
4. D. Generate an SSH key pair in Microsoft Fabric and add the public key to the Azure DevOps user profile.

Wrong-answer review

• A. Configure the Git connection using a Personal Access Token (PAT) generated by the workspace administrator.: Configure the Git connection using a Personal Access Token (PAT) generated by the workspace administrator. is incorrect here because it conflicts with the supported Fabric approach for this scenario. Microsoft Fabric Git integration supports authenticating using a Service Principal. This method allows organizations to connect workspaces to Azure DevOps repositories without relying on individual user accounts. To implement this, the service principal must be registered in Microsoft Entra ID, added as a user in Azure DevOps with appropriate repository permissions, and its application ID and secret configured in the Fabric workspace Git settings. Personal Access Tokens (PATs) and SSH keys are tied to individual user profiles, which violates the requirement to avoid individual accounts. Tenant-level SSO still routes access through the logged-in user's credentials.
• C. Enable Tenant-level Single Sign-On (SSO) and delegate access via Azure DevOps organization-level access policies.: Enable Tenant-level Single Sign-On (SSO) and delegate access via Azure DevOps organization-level access policies. is incorrect here because it conflicts with the supported Fabric approach for this scenario. Microsoft Fabric Git integration supports authenticating using a Service Principal. This method allows organizations to connect workspaces to Azure DevOps repositories without relying on individual user accounts. To implement this, the service principal must be registered in Microsoft Entra ID, added as a user in Azure DevOps with appropriate repository permissions, and its application ID and secret configured in the Fabric workspace Git settings. Personal Access Tokens (PATs) and SSH keys are tied to individual user profiles, which violates the requirement to avoid individual accounts. Tenant-level SSO still routes access through the logged-in user's credentials.
• D. Generate an SSH key pair in Microsoft Fabric and add the public key to the Azure DevOps user profile.: Generate an SSH key pair in Microsoft Fabric and add the public key to the Azure DevOps user profile. is incorrect here because it conflicts with the supported Fabric approach for this scenario. Microsoft Fabric Git integration supports authenticating using a Service Principal. This method allows organizations to connect workspaces to Azure DevOps repositories without relying on individual user accounts. To implement this, the service principal must be registered in Microsoft Entra ID, added as a user in Azure DevOps with appropriate repository permissions, and its application ID and secret configured in the Fabric workspace Git settings. Personal Access Tokens (PATs) and SSH keys are tied to individual user profiles, which violates the requirement to avoid individual accounts. Tenant-level SSO still routes access through the logged-in user's credentials.

Answer choices

1. A. Configure the default Spark pool with a small number of compute nodes and auto-scaling enabled, setting the minimum nodes to 1 and the maximum to 5.
2. B. Configure the default Spark pool with a large number of compute nodes (e.g., 20) to ensure consistent performance regardless of workload.
3. C. Configure the default Spark pool with a medium number of compute nodes (e.g., 8) and disable auto-scaling to maintain predictable resource allocation.
4. D. Configure the default Spark pool with a small number of compute nodes (e.g., 2) and disable auto-scaling to minimize costs during periods of low activity.

Wrong-answer review

• B. Configure the default Spark pool with a large number of compute nodes (e.g., 20) to ensure consistent performance regardless of workload.: Configure the default Spark pool with a large number of compute nodes (e.g., 20) to ensure consistent performance regardless of workload. is incorrect here because it conflicts with the supported Fabric approach for this scenario. The optimal solution is to configure the default Spark pool with auto-scaling enabled and a reasonable range of nodes. Auto-scaling allows the pool to dynamically adjust the number of compute nodes based on workload demands. Setting a minimum of 1 ensures some resources are always available for immediate responsiveness, while a maximum of 5 provides a cost-effective upper limit during peak periods. This approach balances performance and cost by scaling up when needed and scaling down when demand decreases, avoiding unnecessary resource consumption.
• C. Configure the default Spark pool with a medium number of compute nodes (e.g., 8) and disable auto-scaling to maintain predictable resource allocation.: Configure the default Spark pool with a medium number of compute nodes (e.g., 8) and disable auto-scaling to maintain predictable resource allocation. is incorrect here because it conflicts with the supported Fabric approach for this scenario. The optimal solution is to configure the default Spark pool with auto-scaling enabled and a reasonable range of nodes. Auto-scaling allows the pool to dynamically adjust the number of compute nodes based on workload demands. Setting a minimum of 1 ensures some resources are always available for immediate responsiveness, while a maximum of 5 provides a cost-effective upper limit during peak periods. This approach balances performance and cost by scaling up when needed and scaling down when demand decreases, avoiding unnecessary resource consumption.
• D. Configure the default Spark pool with a small number of compute nodes (e.g., 2) and disable auto-scaling to minimize costs during periods of low activity.: Configure the default Spark pool with a small number of compute nodes (e.g., 2) and disable auto-scaling to minimize costs during periods of low activity. is incorrect here because it conflicts with the supported Fabric approach for this scenario. The optimal solution is to configure the default Spark pool with auto-scaling enabled and a reasonable range of nodes. Auto-scaling allows the pool to dynamically adjust the number of compute nodes based on workload demands. Setting a minimum of 1 ensures some resources are always available for immediate responsiveness, while a maximum of 5 provides a cost-effective upper limit during peak periods. This approach balances performance and cost by scaling up when needed and scaling down when demand decreases, avoiding unnecessary resource consumption.

Answer choices

1. A. Read
2. B. ReadAll
3. C. Write
4. D. Contributor

Wrong-answer review

• A. Read: Read is incorrect here because it conflicts with the supported Fabric approach for this scenario. The 'Write' permission level is the appropriate choice for this scenario. It allows users to modify existing items, which aligns with Sarah's need to make adjustments to the dataflow. 'Read' only allows viewing, 'ReadAll' grants access to all items in the workspace, and 'Contributor' provides broader permissions than necessary, potentially allowing Sarah to modify items she shouldn't. The principle of least privilege dictates granting only the permissions required to perform a specific task.
• B. ReadAll: ReadAll is incorrect here because it conflicts with the supported Fabric approach for this scenario. The 'Write' permission level is the appropriate choice for this scenario. It allows users to modify existing items, which aligns with Sarah's need to make adjustments to the dataflow. 'Read' only allows viewing, 'ReadAll' grants access to all items in the workspace, and 'Contributor' provides broader permissions than necessary, potentially allowing Sarah to modify items she shouldn't. The principle of least privilege dictates granting only the permissions required to perform a specific task.
• D. Contributor: Contributor is incorrect here because it conflicts with the supported Fabric approach for this scenario. The 'Write' permission level is the appropriate choice for this scenario. It allows users to modify existing items, which aligns with Sarah's need to make adjustments to the dataflow. 'Read' only allows viewing, 'ReadAll' grants access to all items in the workspace, and 'Contributor' provides broader permissions than necessary, potentially allowing Sarah to modify items she shouldn't. The principle of least privilege dictates granting only the permissions required to perform a specific task.