Compute | Recommendations

Recommendations lists the cost and performance suggestions ADOC generates for your Snowflake and Databricks data sources, based on observed usage and configuration.

Why this matters

Sizing a warehouse or cluster correctly usually isn't a one-time decision — usage patterns shift, and a configuration that made sense at setup can become wasteful or, less often, undersized over time. Recommendations surfaces these opportunities automatically, so you don't have to review every data source manually to find them.

Recommendations expire automatically once they're no longer useful, so the list stays current instead of accumulating outdated suggestions.

Navigating to Recommendations

Navigate to Compute > Recommendations. You can also reach recommendations from the Home page's Recommendations tab.

Reviewing Recommendations

Recommendations are grouped by data source name.

Field	Description
Monitor Name	The monitor associated with the recommendation.
Datasource Name	The data source the recommendation applies to.
Total	The number of clusters included.
Last Evaluation	When the recommendation was last reassessed.

Select the expand control on a group to see the related data sources and inputs in a table. Select a Datasource Name to open the Recommendation Details page for that recommendation.

Recommendation details

Field	Description
Datasource Name	The data source the recommendation applies to.
Monitor Name	The monitor that triggered the recommendation.
Warehouse Name	The warehouse the recommendation is for.
Recommended Warehouse Size	ADOC's suggested size, based on current usage and performance.
Annual Potential Savings	The estimated annual cost savings if you implement the recommendation.
Last Evaluated	When the recommendation was last reassessed.
Action	Applies the recommended change directly from this page.

The fields shown vary depending on the type of recommendation.

By default, only active clusters are shown. Turn off the toggle to view dormant clusters instead.

Exporting Recommendations

Select the export control to download the currently filtered set of recommendations as a CSV file for offline analysis or sharing.

Filtering Recommendations

You can filter recommendations by:

Datasource Type – Narrow to specific data source types, such as Snowflake or Databricks. You can search within the list of types.
Datasource – Narrow to a specific data source by name.

What's next

Compute | Chargeback & Budget – Track whether implementing a recommendation moves spend within budget.
Compute | Monitors – Review the monitor that triggered a given recommendation.