Capital One

How do I know that the Snowflake recommendations applied are actually saving money?

The value report contains a detailed history of cost savings generated by Slingshot recommendations. The cost savings can be viewed by all recommendations as well as per recommendation applied. Additionally, the warehouse details page visualizes the cost over time and will include a change history so that you can see why there were cost changes.

What types of Snowflake cost optimizations does Slingshot cover?

Slingshot optimizes the following objects in Snowflake: Schedules for warehouse size, cluster size and auto-suspend Warehouse statement timeout Just-right time travel provisioning Ideal COPY file sizes Drop unused tables

How are warehouse schedule recommendations formulated?

Slingshot analyzes the workload and query attributes of a warehouse. It tracks factors such as, but not limited to, query size, query load, queued queries and data spillage. Based on these usage patterns, Slingshot generates recommendations that will optimize the Snowflake resources.

Do recommendations support Warehouses with QAS?

Yes, we analyze warehouses that have Query Acceleration Services (QAS) enabled and provide recommendations. Additionally, in some cases, we can enhance QAS cost-efficiency by adjusting the time slots in the scheduling recommendations to account for delayed queries or queries that don't meet the eligibility criteria.

Can the warehouse recommendations impact performance of the jobs running on them?

Slingshot recommendations are designed to be conservative with minimal projected impact to the analyzed workloads. We understand that workloads are not constant, so we give you the tools to manage performance. For example, you can assess each recommendation's impact before applying one, and you can set up a customizable alerting feature to get notified about any SLA breach for thresholds that you define to monitor the performance after recommendation application.

Where can I find requests to provision Snowflake warehouses?

The “Approvals Inbox” contains all warehouse requests—creation, drop, modification and management requests. Requests are accessible from the left nav and include pending, approved and declined requests. The Approvals Inbox is accessible from the home page as well as the left nav and includes open, in progress and a history of closed (approved and declined) requests. Only users with a role of Snowflake Owner, Snowflake Admin or Tenant Admin can submit requests. Approval levels are done on an org-level basis, requiring 1 or 2 levels of approvals. These approval levels are set using approver role

How long is the customer metadata retained in Slingshot?

The default retention period for Capital One Software Customer Data is two years from data creation date. If you have specific data retention policies that are truly applicable to the metadata, please speak with our representatives.

What kind of Snowflake Warehouse information does Slingshot show?

Slingshot shows current and historical cost, performance and usage information for warehouse queries, as well as any performance alerts and available optimizations. For each recommended optimization, Slingshot shows why it is being recommended and what the predicted cost and performance impacts are. Slingshot also provides a detailed snapshot of warehouses. The reporting can be customized to do a cost breakdown analysis or chargeback reporting by using tags.