Need Help Writing a Rule?
SQL Billing Rule Datasets
SQL billing rules comprise two different datasets:Rules written against
aws only affect how data appears in Vantage and do not modify the original CUR file.AWS Dataset
For theaws dataset, you can query a supported subset of AWS CUR columns. Use the Available Columns list in the billing rule editor to see all queryable fields. Custom resource_tag_% columns that vary by account are not supported, though a small set of predefined resource tag columns are available.
Expand the section below for the full list of supported aws columns, organized by CUR table. Use the format aws.table/ColumnName in your SQL (e.g., aws.lineItem/UnblendedCost). Quotes around the column name are optional; the editor accepts both aws.lineItem/UnblendedCost and aws."lineItem/UnblendedCost".
Click to view all supported AWS columns
Click to view all supported AWS columns
bill:
costCategory:
pricing:
product:
reservation:
BillingEntity, BillingPeriodEndDate, BillingPeriodStartDate, BillType, InvoiceId, InvoicingEntity, PayerAccountIdbill/BillingEntity values (from AWS docs)
bill/BillingEntity values (from AWS docs)
The
bill/BillingEntity field identifies whether a charge is for an AWS service or an AWS Marketplace purchase. These values are defined by AWS in the CUR Billing Details documentation.bill/BillType values (from AWS docs)
bill/BillType values (from AWS docs)
The
bill/BillType field indicates the type of bill that the line item covers. These values are defined by AWS in the CUR Billing Details documentation.Project, Team, Environmentdiscount: BundledDiscount, TotalDiscountidentity: LineItemId, TimeIntervallineItem: AvailabilityZone, BlendedCost, BlendedRate, CurrencyCode, LegalEntity, LineItemDescription, LineItemType, NetUnblendedRate, NormalizationFactor, NormalizedUsageAmount, Operation, ProductCode, ResourceId, TaxType, UnblendedCost, UnblendedRate, UsageAccountId, UsageAmount, UsageEndDate, UsageStartDate, UsageTypelineItem/ProductCode common values
lineItem/ProductCode common values
The
lineItem/ProductCode field contains the AWS service code for the line item. The table below lists commonly used product codes. This is not an exhaustive list. AWS product codes can change as services are added or renamed.This is not an exhaustive list. Verify product codes against the AWS Data Exports Data Dictionary or the product code values in your CUR data.
lineItem/LineItemType values (from AWS docs)
lineItem/LineItemType values (from AWS docs)
The
lineItem/LineItemType field indicates the type of charge on each CUR row. These values are defined by AWS in the CUR Line Item Details documentation.LeaseContractLength, OfferingClass, publicOnDemandCost, publicOnDemandRate, PurchaseOption, RateCode, RateId, term, unitpricing/term values (from AWS docs)
pricing/term values (from AWS docs)
The
pricing/term field indicates whether usage is On-Demand or Reserved. These values are defined by AWS in the CUR Pricing Details documentation. This field is empty for Spot Instances, Savings Plan rows, and Marketplace purchases.pricing/PurchaseOption values (from AWS docs)
pricing/PurchaseOption values (from AWS docs)
The
pricing/PurchaseOption field indicates the payment model for a Reserved Instance. These values are defined by AWS in the CUR Pricing Details documentation.APICalls, attachmentType, availability, cacheType, capacitystatus, clockspeed, component, databaseedition, dataTransfer, dedicatedEbsThroughput, deploymentoption, description, destinationCountryISOCode, directconnectlocation, directorysize, directorytype, directorytypedescription, disableactivationconfirmationemail, durability, ebsOptimized, ecu, endpointtype, enhancedNetworkingSupported, filesystemtype, findingGroup, findingSource, freeUsageIncluded, fromLocation, fromRegionCode, fromLocationType, gpu, gpuMemory, group, groupdescription, insightsType, instance, instanceFamily, instanceSize, instanceType, instanceTypeFamily, integratingApi, integratingService, intelAvxAvailable, intelAvx2Available, intelTurboAvailable, invocation, licenseModel, location, locationType, logsDestination, maxIopsBurstPerformance, maxIopsvolume, maxThroughputvolume, memory, messageCountfee, messageType, networkPerformance, normalizationSizeFactor, operatingSystem, operation, originationIdType, osType, parameterType, physicalCores, physicalProcessor, platoClassificationType, pricingUnit, primaryplaceofuse, processorArchitecture, processorFeatures, ProductFamily, ProductName, productSchemaDescription, provisioned, provisioningType, PurchaseOption, purchaseterm, region, regioncode, replicationType, resourceAssessment, resourcePriceGroup, routeType, servicecode, servicename, singleOrDualPass, sizeFlex, sku, storage, storageclass, storagemedia, storagetype, tenancy, throughputCapacity, tier, toLocation, toLocationType, toRegionCode, transcodingResult, trialProduct, upfrontCommitment, usagetype, vcpu, videoCodec, videoFrameRate, videoQualitySetting, volumeType, workforceTypeproduct/ProductFamily common values
product/ProductFamily common values
The
product/ProductFamily field categorizes the type of product. The table below lists commonly seen values. This is not an exhaustive list; product family values vary by service and can change over time.This is not an exhaustive list. Product family values vary by service and can change over time. Verify against the AWS Data Exports Data Dictionary or the values in your CUR data.
AmortizedUpfrontCostForUsage, AmortizedUpfrontFeeForBillingPeriod, AvailabilityZone, EffectiveCost, EndTime, ModificationStatus, NetAmortizedUpfrontCostForUsage, NetAmortizedUpfrontFeeForBillingPeriod, NetEffectiveCost, NetRecurringFeeForUsage, NetUnusedAmortizedUpfrontFeeForBillingPeriod, NetUnusedRecurringFee, NetUpfrontValue, NormalizedUnitsPerReservation, NumberOfReservations, RecurringFeeForUsage, ReservationARN, StartTime, SubscriptionId, TotalReservedNormalizedUnits, TotalReservedUnits, UnitsPerReservation, UnusedAmortizedUpfrontFeeForBillingPeriod, UnusedNormalizedUnitQuantity, UnusedQuantity, UnusedRecurringFee, UpfrontValueresourceTags: user:creator, user:name, user:owner, user:purposesavingsPlan: AmortizedUpfrontCommitmentForBillingPeriod, EndTime, InstanceTypeFamily, NetAmortizedUpfrontCommitmentForBillingPeriod, NetRecurringCommitmentForBillingPeriod, NetSavingsPlanEffectiveCost, OfferingType, PaymentOption, PurchaseTerm, RecurringCommitmentForBillingPeriod, Region, SavingsPlanARN, SavingsPlanEffectiveCost, SavingsPlanRate, StartTime, TotalCommitmentToDate, UsedCommitmentsplitLineItem: ActualUsage, NetSplitCost, NetUnusedCost, ParentResourceId, PublicOnDemandSplitCost, PublicOnDemandUnusedCost, SplitCost, SplitUsage, SplitUsageRatio, UnusedCostCosts Dataset
The table below is the full data dictionary of allcosts columns you can use in billing rule SQL.
Supported Cost Type Values
Thecost_type column corresponds to the Charge Type filter in Cost Reports. Expand the section below for the exact raw costs.cost_type values you can use in SQL.
Click to view all supported cost_type values
Click to view all supported cost_type values
Each AWS discount type in the table above, such as
BundledDiscount, EdpDiscount, and PrivateRateDiscount, can also have amortized and unamortized variants that use the raw costs.cost_type value with a prefix. For example, EdpDiscount can also appear as AmortizedEdpDiscount or UnamortizedEdpDiscount.- Amortized variants such as
AmortizedEdpDiscountappear when amortization and discounts are both enabled. These represent the discount amount on fee rows, such as Marketplace or Enterprise Support charges, that has been spread across the billing period. - Unamortized variants such as
UnamortizedEdpDiscountappear when amortization is disabled and discounts are included. These represent the discount amount on unamortized fee or commitment rows, such as Reserved Instance fees or Savings Plan recurring fees.
EdpDiscount, BundledDiscount, SppDiscount, RiVolumeDiscount, PrivateRateDiscount, and DistributorDiscount.Supported Provider Values
Thecosts.provider field uses lowercase provider codes. Use these exact values when filtering by provider in your SQL.
Click to view all supported provider values
Click to view all supported provider values
Azure-Specific Field Mapping
For Azure,costs.provider_account_id and costs.resource_account_id map to different Azure identifiers depending on your workspace configuration:
Some workspaces use a Vantage setting that swaps these two columns. If your Azure rule is not matching the expected rows, compare the values in your rule against a known cost line (via CSV export or API) to confirm which column holds the subscription for your workspace.
costs to aws Column Mapping
When Vantage processes AWS CUR data, it normalizes CUR columns into the costs schema. If you are choosing between writing a rule against the aws dataset (raw CUR) or the costs dataset (normalized), the table below shows how the primary CUR fields map to costs columns.
Click to view costs to AWS CUR column mapping
Click to view costs to AWS CUR column mapping
Some CUR columns like
pricing/publicOnDemandCost and pricing/publicOnDemandRate are only available in the aws dataset. If your rule needs to reference on-demand pricing (e.g., for RI/SP re-rating), use the aws dataset instead of costs.SQL Syntax Reference
This section starts with the basic structure for SQL billing rules: which SQL operations are supported, how to format column names, which operators and expressions you can use, and how Monthly Insert aggregate rules work.SQL Operations
Use the following SQL operations to define billing rules:Column Name Format
All column references must be fully qualified with the dataset name:costsdataset: Usecosts.column_nameformat (e.g.,costs.amount,costs.service)awsdataset: Useaws.table/ColumnNameformat (e.g.,aws.lineItem/UnblendedCost,aws.bill/BillingEntity)
Supported Operators
The following operators and expressions are supported inWHERE clauses and SET assignments:
You can include leading SQL comments before the rule statement. Both
-- line comments and /* */ block comments are supported.
Monthly Insert Aggregate Rules
Monthly Insert aggregate rules useINSERT INTO costs ... SELECT ... FROM costs to calculate a value from existing cost rows and insert a new monthly cost row. Use them when the rule needs the total of a set of costs, such as a management fee, surcharge, or tiered support fee.
These rules run after the other billing rule stages have processed. They read from the transformed costs dataset, so prior markups, discounts, inserted charges, and exclusions can change the total that the aggregate rule sees. See Billing Rule Execution Order for the full stage model.
Monthly Insert Basic Structure
SUM(), AVG(), MIN(), MAX(), and COUNT(). You can use CASE WHEN inside the SELECT to define conditional logic, and you can use GROUP BY and HAVING when the inserted rows need to be grouped.
A Monthly Insert aggregate rule must include at least one aggregate function in the SELECT. A rule without one is rejected, since an INSERT ... SELECT that does not aggregate is likely not the calculation you intended.
Monthly Insert aggregate rules are available only with the
costs dataset. Use INSERT INTO costs ... SELECT ... FROM costs; do not use INSERT INTO aws or a standalone SELECT.Cost Settings
The Cost Settings dropdown appears on SQL billing rules, but it only affects Monthly Insert aggregate rules that calculate from existingcosts rows. Changing these settings on row-level UPDATE, DELETE, or fixed VALUES INSERT rules does not change how those rules run.
For Monthly Insert aggregate rules, Cost Settings control which source rows are available to the aggregate calculation before your SQL WHERE clause is applied:
Cost Settings and your SQL filters work together. For example, if Credits are excluded in Cost Settings, credit rows are not part of
SUM(costs.amount) even if the SQL does not explicitly filter them out. If amortization is disabled, one-time upfront charges, such as upfront Reserved Instance and Savings Plan fees, are included as they appear.
Examples:
- If your rule calculates
SUM(costs.amount) * 0.10and Credits are excluded, the 10% fee is calculated before credits reduce the total. - If Taxes are included, tax rows are part of the total unless your SQL also filters them out, such as
AND costs.cost_type != 'Tax'. - If Amortization is disabled, a one-time upfront Reserved Instance or Savings Plan fee can be included in the month it appears. If Amortization is enabled, the fee is calculated from amortized cost rows instead.
Monthly Date Behavior
When a Monthly Insert aggregate rule omits thedate column, Vantage derives the date from the source cost rows and creates one inserted row for each month of qualifying data. For a single-month period, this produces one row. If the source costs span multiple months, the rule produces one row per month.
The range of months a rule can produce is bounded by the rule’s Start Date and End Date, which you set in the rule configuration rather than in the SQL. Use those fields to scope the rule to a specific period.
Examples:
- If the qualifying costs are all from January, the rule inserts one January fee row.
- If the qualifying costs include January and February rows, the rule inserts one January fee row and one February fee row.
- If you set the rule’s Start Date and End Date to a single month, the rule inserts a row only for that month.
How to Construct SQL Statements
Step-by-step guides for building UPDATE, DELETE, and INSERT statements.UPDATE Statements
UPDATE Basic Structure
1
Start with UPDATE
Choose your dataset (
costs or aws).2
Specify what to change
Use
SET to define the new value. You can update any column from the costs data dictionary or AWS column list, not just amount.3
Add conditions
Use
WHERE to specify which rows to modify.- Apply discount:
SET costs.amount = costs.amount * 0.90(10% discount) - Apply markup:
SET costs.amount = costs.amount * 1.15(15% markup) - Set fixed amount:
SET costs.amount = 100.00
DELETE Statements
DELETE Basic Structure
1
Start with DELETE FROM
Choose your dataset (
costs or aws).2
Add conditions
Use
WHERE to specify which rows to remove.- Remove by service:
WHERE costs.service LIKE 'AWS Support%' - Remove by cost type:
WHERE costs.cost_type = 'Fee' - Remove by amount:
WHERE costs.amount < 0.01 - Remove by multiple conditions:
WHERE costs.provider = 'aws' AND costs.service = 'Amazon S3'
INSERT Statements
INSERT statements can only be used with the
costs dataset. INSERT is not supported for the aws dataset.INSERT Basic Structure
VALUES when adding a fixed charge or credit. Values must be static literals; you cannot reference columns or use expressions (e.g., costs.amount * 0.10) in the VALUES clause.
Use INSERT ... SELECT for Monthly Insert aggregate rules that calculate a value from existing costs:
Monthly Insert Basic Structure
cost_type you insert must be one of the supported cost_type values, and any provider you filter on or insert must be one of the supported provider values.
1
Start with INSERT INTO
Use the
costs dataset for INSERT statements.2
List required columns
Specify all necessary columns.
3
Add VALUES
Provide values for each column.
- Add credit:
VALUES ('aws', 'Billing Credit', 'Billing Credit', 'Billing Credit', 'Credit', -100) - Add processing fee:
VALUES ('aws', 'Processing Fee', 'Processing Fee', 'Processing Fee', 'Fee', 5) - Add percentage fee on total spend:
SELECT SUM(costs.amount) * 0.10, ... FROM costs WHERE ...
Create SQL Billing Rules
To create a new SQL billing rule:1
Create a new custom rule
- From the top navigation, click Settings.
- From the left navigation, under Partner Settings, select Billing Rules.
- Click New Billing Rule, then select Custom.
2
Configure the rule
Add the following rule details:
- Title: Give your rule a descriptive name.
- Start Date/End Date: Define the date range for which the rule should apply. Use these fields instead of date logic in your SQL. If you leave both blank, the rule applies to all billing periods (past and future). You can set only a Start Date (applies from that date forward) or only an End Date (applies up to that date).
- Select whether you want to automatically apply this billing rule to all existing and new managed accounts.
3
Write the rule SQL
- From the Available Columns list, select either the AWS/
awsor Vantage/costsschema. The selected schema must match the dataset in your SQL. If you write acostsrule but leave the schema set toaws, saving will fail. - Under Cost Fields, view available fields and click to copy, or type any column from the costs data dictionary or AWS column list directly into the editor.
- In the SQL editor, write your SQL statement using the selected dataset and columns. The editor supports autocomplete and syntax highlighting.
- Click Format above the editor to auto-format your SQL with consistent indentation and line breaks.
- Click Validate to check your query against the selected schema before saving. Validation catches syntax errors, unsupported operations, schema mismatches (e.g., using
costscolumns against theawsdataset), and common issues like missing spaces before SQL keywords or multiple statements in a single rule.
- Click Submit to create the rule.
After You Save a Rule
Once you submit a rule, keep the following in mind about how it is processed and how it interacts with your other rules.- Reprocessing: When you save a billing rule, Vantage automatically reprocesses your cost data back to the rule’s start date. If you leave the start date empty, the rule applies to your entire history, which means more data to reprocess. Reprocessing can take several hours depending on the size of the account, and saving multiple times restarts the cycle. Finalize your SQL before saving, and test after reprocessing completes to confirm the result.
- Rule limit: Each workspace can define up to 20 billing rules. If you need more, contact support@vantage.sh. For multiple adjustments with different conditions, create separate rules.
-
Monthly Insert rules: Aggregate
INSERT ... SELECTrules run in the Monthly Insert stage, after line-item transform rules. This lets a rule calculate a fee from the final monthly cost set after prior markups, discounts, exclusions, or inserted charges have been applied. -
Rule order: Rules are applied sequentially, so order matters when multiple rules touch the same line items. Each statement operates on the data left behind by the rules before it.
For example, consider two rules that both target the same Platform Fee line item—one that sets a fixed amount and one that applies a 10% discount:
Rule A — set a fixed amountThe final amount depends on which rule runs first:Rule B — apply a 10% discount
- A then B: the amount is set to
$100, then discounted to$90. - B then A: the amount is discounted first, then overwritten to
$100—the discount is lost.
- A then B: the amount is set to
Example Rules
Discounts and Markups
Apply a 5% Discount (AWS)
Apply a 5% discount to all cost line items, excluding credits:At this time, multiplication is supported, but not division.
Apply Service-Specific Markup
Apply a 5% markup to specific services:Rename a Service
Rename a service label in normalized cost data:Apply Markup by Cost Category
Apply a 10% markup to all storage costs:Reduce Costs for a Specific AWS Account
Apply a 50% cost reduction to a specific linked account:Set Credits to Zero
Zero out all credit line items:Zero Out Cost and Rate for Internal Accounts
Set both the cost and rate to zero for specific internal accounts (e.g., cost optimization or internal operations accounts):Exclusions
Exclude SPP Discounts
Remove Solution Provider Program discounts from billing:Exclude Multiple SPP Discount Types
Remove both SPP and unamortized SPP discounts:Exclude SPP/EDP Discount Line Items (AWS)
Remove Enterprise Discount Program and Solution Provider discount or credit line items at the CUR level:Exclude AWS Support Products (AWS)
Remove all AWS Support product costs at the CUR level:Exclude Support Costs
Remove AWS Business support costs from billing:Exclude Marketplace Costs
Remove all AWS Marketplace costs from billing:Exclude Marketplace by Billing Entity (AWS)
Remove AWS Marketplace costs at the CUR level using the billing entity field:Exclude BundledDiscount Line Items (AWS)
Remove BundledDiscount line items from the CUR:Remove Costs for Specific AWS Accounts
Exclude all costs from specific linked AWS accounts:Account-based exclusions must use account IDs, not account names. The AWS CUR does not include an account name column that can be used in billing rules.
Exclude Other Provider Items
Remove Temporal Cloud Actions from billing:Reserved Instance and Savings Plan Adjustments
Unamortize Reserved Instance and Savings Plan Costs
Convert amortized Reserved Instance and Savings Plan costs back to usage-based costs. This rule sets the line item type toUsage when the Savings Plan or reservation does not belong to the linked account (using string concatenation to compare account IDs within ARNs):
Unamortize Savings Plans for a Specific Account
Convert Savings Plan covered usage back to on-demand for a specific account by matching the account ID within the Savings Plan ARN:Rerate Reserved Instance Costs to On-Demand Pricing
Replace discounted Reserved Instance costs with public on-demand pricing by copying values from thepricing columns. This sets both the cost and rate to their on-demand equivalents for reservations belonging to specific accounts:
When filtering by account ID within an ARN field, use
LIKE '%account_id%' to match the account ID substring within the full ARN string. Do not use IN with bare account IDs — ARN fields contain full ARN strings (e.g., arn:aws:ec2:us-east-1:111111111111:reserved-instances/...), so an IN comparison against just '111111111111' will never match.Exclude Reservation Fees for Specific Accounts
Remove Reserved Instance fees that belong to specific internal accounts without affecting other usage on those accounts. UseLIKE to match the account ID within the reservation ARN:
Monthly Insert Aggregate Rule Examples
Monthly Insert aggregate rules use
INSERT INTO costs ... SELECT ... FROM costs. They calculate from existing monthly cost rows and insert new synthetic cost rows.Add a 10% Management Fee on Qualifying AWS Spend
Insert a management fee equal to 10% of AWS spend, excluding Marketplace purchases and EDP/SPP discount line items:Add a Tiered Support Fee
UseCASE WHEN around SUM() when the fee rate changes based on the customer’s total qualifying spend. This example charges 10% on the first $10,000, 7% on the next $70,000, and 5% above $80,000:
Insert Line Items
INSERT statements can only be used with the
costs dataset.Add a Monthly Charge
Insert a flat monthly charge as a new line item:Add a Credit
Insert a billing credit as a negative amount:Add a Processing Fee
Insert a processing fee for services rendered:Add a Support Fee with a Specific Date
Insert a support fee for a specific billing period. You can include thedate column to specify which month the charge applies to:
Azure Examples
Azure billing rules use the
costs dataset. There is no separate Azure CUR dataset. When writing Azure rules, use costs.provider = 'azure' for standard Azure or costs.provider = 'azure_csp' for Azure CSP. These are distinct provider values, and rules must target the correct one.Apply Markup to Azure Services
Apply a 10% markup to Azure Virtual Machines:Apply Discount Excluding Reservations (Azure)
Apply a discount to all Azure costs while excluding reservation purchases:Apply Markup by Azure Subscription
Apply a 10% markup to costs from a specific Azure subscription. See the Azure-Specific Field Mapping section to determine whethercosts.resource_account_id or costs.provider_account_id contains the subscription ID for your workspace.
Exclude Partner Earned Credit (Azure CSP)
Remove Partner Earned Credits from Azure CSP billing. Note the use ofazure_csp as the provider value: