Skip to main content

Keebo Frequently Asked Questions

Warehouse Optimization for Snowflake​

Does Warehouse Optimization Support Snowpark-Optimized Warehouses?​

Warehouse Optimization is available for Snowpark-optimized warehouses. The smallest allocatable instance type for these warehouses is Medium.

What SSO Providers Are Supported?​

Keebo supports any identity provider that supports OIDC for SSO.

How Does Warehouse Optimization Account for Optimizations That Increase Query Costs?​

Warehouse Optimization does not count negative savings impact. If an optimization slows queries enough to increase costs, savings are rounded to 0 rather than reported as a negative value.

Keebo supports connecting to Warehouse Optimization with an AWS private link. Mention this during onboarding or contact Keebo support for more information.

How Does Warehouse Optimization Calculate Savings?​

Savings calculation details are available by clicking "Savings Calculations" in the Saved Credits chart on the Overview page. Contact Keebo support for additional information.

Why Does Keebo Run SHOW WAREHOUSE or SELECT 1 Commands Frequently?​

SHOW WAREHOUSE commands determine whether a warehouse is running or idle. This information is used to decide whether to suspend a warehouse. SELECT 1 queries are issued by the standard JDBC driver used to connect to Snowflake and help manage connection pools.

Do SHOW WAREHOUSE or SELECT 1 Commands Cost Snowflake Credits?​

Neither command uses Snowflake compute credits. Both run in the Cloud Services layer. Keebo does not wake up warehouses for these commands.

What Time Zone Are Dashboard Times Displayed In?​

All charts display times in the UTC time zone. The time zone cannot currently be changed.

Warehouse Optimization for Databricks​

Preview

Warehouse Optimization for Databricks is currently in preview. Reach out to Keebo support for access and onboarding.

How Does Warehouse Optimization Connect to a Databricks Environment?​

Warehouse Optimization connects using OAuth Machine-to-Machine (M2M) authentication via a service principal. It uses the Databricks REST API to retrieve warehouse metadata and apply real-time optimizations, and queries Databricks system tables to calculate savings. Keebo never accesses organization data directly.

What Data Does Warehouse Optimization Access?​

Warehouse Optimization accesses only the metadata necessary for optimization. Via the REST API, it retrieves warehouse configuration, applies warehouse configuration changes, and reads recent query history as input to its algorithms. Via system tables, it reads warehouse events, warehouse configuration changes, billable usage data, and query history. Every piece of data retrieved serves a specific role in the optimization process.

How Long Does It Take to See Savings After Connecting Warehouses?​

After connecting a Databricks workspace and SQL warehouses, Warehouse Optimization begins optimizing immediately. Savings are typically reflected in the dashboard within 24 hours.

Can Warehouse Optimization Manage Multiple Databricks Workspaces?​

Warehouse Optimization supports multiple workspaces and by default uses a single service principal across all connected workspaces. Contact Keebo support if separate service principals per workspace are needed.

What Permissions Does the Keebo Service Principal Need?​

The service principal requires:

  • User-level access to each Databricks workspace
  • Access to Databricks system tables (granted via SQL)
  • Read/write access to the Keebo catalog and schema
  • CAN_MANAGE permission on each SQL warehouse to be optimized

Can Warehouse Optimization Be Disabled for a Specific Warehouse?​

All Warehouse Optimization optimizations can be turned off for any warehouse at any time using the toggle in the "Keebo status" column on the Settings page.

What Happens if a Warehouse's Settings Are Changed Outside of Warehouse Optimization?​

Warehouse Optimization automatically detects changes made outside of Keebo to connected warehouses' size, auto-stop, and cluster values. By default, Warehouse Optimization continues optimizing after a change is detected, treating the new setting as the default. This behavior can be configured in the warehouse settings.

Does Warehouse Optimization Support Classic and Pro SQL Warehouses?​

Warehouse Optimization is currently only works for Databricks Serverless SQL Warehouses. Contact Keebo support for questions about other warehouse types.

Workload Intelligence​

Keebo supports connecting to Workload Intelligence with an AWS private link. Mention this during onboarding or contact Keebo customer success for more information.

Does Workload Intelligence Check for Snowpipe Activities When Determining Unused or Unread Tables?​

Snowpipe is a wrapper over the COPY INTO command. Snowflake records COPY commands in the ACCOUNT_USAGE.ACCESS_HISTORY view that Workload Intelligence uses to retrieve table access and usage details.

Can the Existing Warehouse Optimization Schema and Views Be Reused for Workload Intelligence?​

Reusing schemas and views across different products is not recommended. Sharing a schema or view structure can introduce unintended dependencies, where changes in one product could impact the functionality of another. To ensure stability and flexibility, create separate schemas and views tailored to each product's requirements.

Does the Workload Intelligence Warehouse Need to Be Medium?​

A Medium warehouse size is recommended to ensure reasonable performance during data processing and loading. Organizations with fewer queries can experiment with a Small warehouse, while organizations with a large amount of Snowflake usage may benefit from a larger warehouse. Using a warehouse that is too small can result in data failing to load.

How Much Does the Workload Intelligence Warehouse Cost?​

The cost depends on the warehouse size and the queries run. For most organizations, the cost is minimal. Predicting the exact cost before onboarding is not possible, as it varies with usage patterns and query demands.

How Can Query Text Be Protected From Keebo Access?​

The QUERY_HISTORY view can be modified to replace the QUERY_TEXT column with a hashed version. This ensures query text remains private while still allowing the necessary analysis. Work with Keebo customer success during onboarding to implement query hashing.

Why Is There No Data After Onboarding?​

Pre-aggregation events take approximately one hour to complete. Dashboards remain empty until this process finishes. If no data appears after 24 hours, contact Keebo support.