Limited Release
A metric defines how to aggregate event data and calculate billable usage. An event is an action your customer performs while using your system, such as placing an API call.
Metrics connect the following:
- Pricing plans that apply pricing models to metric usage for calculating charges.
- Usage events that the metric aggregates to determine billable consumption.
1. Decide the event to track
Before you create metrics, choose which event you want to track and the event details (event properties) you require. For most aggregation methods, you need to include the event property name in the API call to create metrics.2. Decide the metric type
Based on your business use case, decide on one of the following metric types:- Metered metric: Customers are charged based on exactly how much of a feature they use within a specific billing period. The usage resets to zero each cycle and customers pay only for what they use in that period. For example, a metric that calculates usage based on the number of API calls placed in a month.
- Recurring metric: Customers are charged for usage that persists across billing periods. The usage accumulates or remains steady until modified. For example, a metric that calculates usage as the total storage space utilized.
3. Decide the aggregation type and aggregation field
Aggregation type determines how the metric aggregates event data and calculates usage. Based on your business use case, choose an aggregation type from the following.| Aggregation type | Description | Use case |
|---|---|---|
COUNT | Counts total items | When you need a simple total, like API calls or transactions. |
COUNT_DISTINCT | Counts unique values in an events data set | When you want to track distinct users, devices, or IDs. |
SUM | Adds up values in a data set | When you want a total amount, such as data usage or time spent. |
MAX | Finds the highest value in a data set | When you need to know the peak value, like maximum storage or highest transaction. |
LATEST | Uses the most recent value in a data set | When only the current or last value matters, like current balance or latest status. |
Recurring metrics support only
SUM and COUNT_DISTINCT aggregation types.storage_gb and region. You can set up a metric that aggregates storage_gb to determine billable usage. In this case, storage_gb is the aggregation field.
You need to pass the aggregation type in the API call to create metrics.
4. Create metrics
Use a valid access token and make a POST call to the/v1/commerce/billing/metrics endpoint. Include the following parameters:
| Parameter | Action |
|---|---|
nameRequired, string | Provide a descriptive name that identifies the metric’s purpose. |
codeRequired, string | Set a unique code to identify this metric. |
typeRequired, string | Set to METERED to reset usage each billing period or RECURRING to accumulate usage. |
aggregation_typeRequired, string | Choose the aggregation type based on your billing use case. See Aggregation types for more information. Possible values: COUNT, COUNT_DISTINCT, SUM, MAX, LATEST. |
aggregation_fieldRequired when aggregation_type is SUM, COUNT_DISTINCT, MAX, or LATEST, string | Set to the event property name that contains the value to aggregate. |
descriptionstring | Provide additional details about what the metric measures. |
field_filters[]array | Add filter objects to include only specific events. See Field filters for more information. |
field_filters[].keyRequired when field_filters[] provided, string | Set to the event property name to filter by. |
field_filters[].valuesRequired when field_filters[] provided, array | Provide an array of values to include in the metric calculation. |
201 Created response. The response includes the following parameters:
| Parameter | Description | Further action |
|---|---|---|
idstring | Unique identifier for the metric. | Use this id when creating pricing plans. |
codestring | Unique code for the metric. | Use this code when managing the metric. |
5. Filter aggregation field values
Field filters let you include only specific events or data in your metric calculation. The system only includes events that match all your filter rules. For each filter, choose a property name (key) and the values you want to match. Common ways to use filters and samples:- By region: Track usage across different geographic locations.
- By performance tier: Separate different service levels.
- By token type: Distinguish between different AI processing types.
- By feature: Track premium vs standard features.
- By status: Count only successful operations.
6. Manage metrics
You can update metric configurations when adjusting aggregation methods, modifying field filters, or changing metric descriptions. To review a metric’s current configuration, call the Get metric details endpoint.| Metric attribute | Updatable | Possible management options | Impact on existing plans |
|---|---|---|---|
| Metric name | Yes | Update metric details - modify display name | No impact - display only |
| Metric description | Yes | Update metric details - modify metric description | No impact - display only |
| Aggregation type | Yes | Update metric details - modify calculation method | Affects usage calculations for all plans using this metric |
| Aggregation field | Yes | Update metric details - modify event property used for aggregation | Affects usage calculations for all plans using this metric |
| Field filters | Yes | Update metric details - modify event filtering rules | Affects which events count toward usage |
| Metric code | No | Metric cannot be modified - unique identifier remains permanent | N/A - cannot be modified |
| Metric type | No | Metric cannot be modified - type (METERED or RECURRING) is permanent | N/A - cannot be modified |