View edition slot recommendations

The BigQuery slot recommender creates recommendations foredition or on-demand workloads. The recommenderanalyzes historical slot usage for query jobs and calculates the cost-optimalsettings for edition commitment slots andautoscaling slots while maintainingsimilar performance. The slot recommender can also recommend a maximumreservation size that can improve performance.

You can use the slot recommender for both reservations billing and on-demandbilling:

  • For reservations billing, you can get a cost-optimized recommendation forEnterprise or Enterprise Plus edition workloads andperformance-based recommendations for your reservations.
  • For on-demand billing, you can get cost-optimized recommendations foron-demand workloads across the organization, for a specific project, ora group of projects if you were to convert one or more projectsto the Enterprise edition.

For more information about the recommender service, see theRecommender overview.

Cost-optimized recommendations

The slot recommender estimates your autoscaling usage based on slot usage overthe past 30 days. For more information about slots autoscaling, seeIntroduction to slots autoscaling.The recommender can generate multiple commitment options and calculate thetotal cost for each option. The recommender can also recommend options with thelowest total cost by using custom pricing. The recommended commitment andautoscale slots are meant to cover P99 slot usage of the entire 30-dayobservation window.

The slot recommender offers recommendations for different pricing types,including pay as you go (no commitments), 1-year and 3-year commitments. Itdisplays the monthly cost for each option based on custom pricing.

Slot usage chart in theGoogle Cloud console.

Slot recommendations in theGoogle Cloud console.

The recommendation includes the following details:

  • Baseline commitment slots: The number of commitment slots to achieve optimalcost without affecting performance. You can also view the optimalcommitments in the usage chart above by selectingView optimalcommitments.
  • Baseline commitment monthly cost: The monthly cost of the optimal commitmentslots, calculated using the custom edition commitment price. A month isdefined as 730 hours.
  • Autoscale slots: The maximum number of autoscale slots used at a time. Thisrepresents the additional slots beyond the optimal commitment slots that arecovered by autoscaling. This value does not include the commitment orbaseline slots.
  • Expected autoscale utilization: The expected monthly utilization ofautoscale slots, calculated as the expected autoscale slots used divided bythe maximum autoscale slots.
  • Autoscale monthly cost: The monthly cost of using the expected amount ofautoscale slots, calculated using the custom autoscale price.
  • Total monthly cost: The total monthly cost, which includes the commitmentmonthly cost and the autoscale monthly cost.

Best practices when applying recommendations

Note: The recommendations provided are based on historical data, and the actualresults may vary depending on the specific characteristics of your workloads.
  1. Ensure that the sum of baseline slots for all reservations under the editionis equal or less than the commitment slots. This ensures that any slot usageexceeding the commitment slots can be covered by autoscale slots. If thebaseline slots exceed the commitment slots, you are billed for additionalbaseline slots.
  2. Autoscale slots in the settings is chosen so that available capacity matchesthe peak historical usage. This is to ensure the performance is notimpacted. You can also adjust the autoscale slots to a value lower than themaximum, which can increase autoscale utilization. However, please note thatwhen your slot usage cannot be fully covered, it may affect queryperformance.
  3. In circumstances where your workload experiences spikes and temporarilyexceeds its maximum capacity, the slot recommender mechanism can overestimate its recommendations. In such situations, you may consider maintaining the current settings, assuming that you are satisfied with the current level of performance.

You might see the messageSlot Estimator doesn't have any recommendationsthat would be more effective than your current settings even if your slot usagesometimes exceeds your configured maximum. This can occur because BigQueryoccasionally overprovisions slots temporarily to boost query speed, at no extracost. The slot recommender aims to maintain your recent performance, includingthese bursts. If this P99 usage is higher than your current maximum, anyrecommendation at or below your current setting wouldn't meet those performancelevels, so no change is suggested.

Required permissions

To view cost-optimal commitment slots recommendations, you need the followingIdentity and Access Management (IAM) permissions:

  • recommender.bigqueryCapacityCommitmentsRecommendations.get
  • recommender.bigqueryCapacityCommitmentsRecommendations.list

Each of the following predefined IAM roles includes thesepermissions:

  • BigQuery Resource Admin
  • BigQuery Slot Recommender Viewer
  • BigQuery Slot Recommender Admin

To view recommendations for your edition workloads, you must have the listedpermission for the administration project.

To view project level recommendations for your on-demand workloads, you musthave the previously listed permissions at the project level.

To view cost-optimized recommendations for on-demand workloads for a group ofprojects, you must have the previously listed permissions at the organizationlevel and eitherbigquery.jobs.listExecutionMetadata orbigquery.jobs.listAll permission.

To view organization level recommendations for your on-demand workloads, youmust have the previously listed permissions at the organization level. You alsoneed theresourcemanager.organizations.get permission. TheOrganizationViewer IAM role includes this permission.

In the recommendations settings, rows such asBaseline commitment slots andTotal monthly cost are visible, butthe values of the monthly cost details are hidden. To view the hidden values,you also need the following permission:

  • billing.accounts.getPricing

Each of the following predefined IAM roles includes thesepermissions:

  • Billing Account Viewer
  • Billing Account Administrator

For edition workloads, you need the permissions listed above at the billingaccount associated with the administrator project. For project level on-demandworkloads, you need the permissions at the billing account associated with theproject or at the organization level for organization level recommendations.

For more information about IAM roles in BigQuery,seePredefined roles and permissions.

Performance-improving recommendation

When you select a specific reservation workload, the slot recommender suggeststhe maximum reservation size that can enhance job performance. The slotestimator then analyzes theslot modeling data andfinds the minimum incremental value for the maximum reservation size that iscapable of elevating job performance by at least 5%. If your current maximumreservation size meets your historical needs, then no recommendation is made.

Slot recommender reservation recommendation in theGoogle Cloud console.

Note: Recommendation are based on historical data. Job performance can varybased on actual use.

To implement a recommendation, clickApply to be redirected to the pagewhere you can update the reservation.

Slot recommender reservation recommendation edit reservation in theGoogle Cloud console.

Before you begin

Before you can view the recommendations, you mustenable the RecommenderAPI. To view the recommendations within theGoogle Cloud console, you must alsoenable the ReservationsAPI.

Required permissions

The slot recommender for reservations performance improvement recommendationsrequires that you have the following IAM permissions on theadministration project:

  • bigquery.reservations.list
  • bigquery.reservationAssignments.list
  • bigquery.capacityCommitments.list

To apply the recommended updates to the reservation, you must also have thefollowing IAM permissions on the administration project:

  • bigquery.reservations.update

For more information about IAM roles in BigQuery,seePredefined roles and permissions.

Pricing

This recommender is displayed within the context ofslot estimator. You can use the recommendations at no charge.

View slot recommendations

To view slot recommendations using the Google Cloud console, perform the followingsteps.

  1. In the Google Cloud console, go to the BigQuery page.

    Go to BigQuery

  2. For edition workloads recommendations, select the administrator project. Foron-demand workloads recommendations, select any project within yourorganization that meets the predefined requirements.

  3. For on-demand workloads, if you have organization level permissions, you canselect any individual project or the entire organization in the side paneloptions to view recommendations for the specific scope.

  4. In the navigation menu, clickCapacity management.

  5. Click theSlot estimator tab.

  6. In theSource pane, select an on-demand workload or anedition (Enterprise or Enterprise Plus) workload.

    • If you select an edition workload, detailed recommendations appear underthe graph of historical usage.
    • If you select an on-demand workload, the organization administrator isable to switch between organization level and projectlevel (for one or more projects).

Except as otherwise noted, the content of this page is licensed under theCreative Commons Attribution 4.0 License, and code samples are licensed under theApache 2.0 License. For details, see theGoogle Developers Site Policies. Java is a registered trademark of Oracle and/or its affiliates.

Last updated 2026-02-18 UTC.