Skip to main content
PRIVATE BETAThis feature is currently in Private Beta, contact sales@metronome.com for early access.Don’t have a Metronome account? Sign up to get started.
Rated events turn raw usage into priced, atomic revenue units by applying each customer’s pricing, products, and credits to every event. They show exactly how much each event costs, powering precise billing, revenue and margin analysis, and customer-facing spend transparency from a single, unified dataset.

How rated events work

When Metronome rates a usage event, it resolves the event against a customer’s active contract, including their products, pricing, and any commits or credits. The result is one or more rated events, each representing a distinct priced unit of revenue. A single usage event can produce multiple rated events in the following situations:
  • Multiple products matched: The event matches more than one billable metric (for example, a request that counts toward both an API calls product and a data transfer product).
  • Commit boundary crossed: Part of the usage is covered by a prepaid commit and the remainder is overage, so the covered portion and the overage portion appear as separate rated events.
  • Multiple credits applied: Different credit types (for example, a promotional credit and a standard commit) each apply to different portions of the usage.
  • Multiple contracts matched: The event matches more than one active contract. Note that this pattern is not a best practice, as it becomes difficult to track consumption correctly across multiple contracts.
Each rated event maps the transaction back to the customer, contract, invoice, product, and any commit or credit applied.

What you can do with rated events

  • Aggregate spend by any dimension included in event properties (region, model, cluster, workspace, user, etc.)
  • Aggregate spend by any time dimension (minute, hour, day, week, month)
  • Combine property dimensions with time dimensions for custom breakdowns
  • Join back to the source events table via transaction_id for full end-to-end traceability

Rated events vs. other data models

Metronome exports several related data models that serve different purposes. Understanding how they differ helps you choose the right dataset for your analysis. Use finalized invoice data when you need final billed amounts and invoice status. Use finalized breakdowns when you need daily totals within a billing period. Use finalized rated events when you need event-level detail, custom time windows, or spend breakdowns by properties such as region, model, cluster, or user.

Use cases

Finance
  • Revenue recognition: Support complex performance obligation calculations, nth-of-month billing, and more.
  • SOX audit: Provide auditors with complete end-to-end auditability from usage event through to paid invoice.
Engineering
  • Margin optimization: Aggregate spend by compute cluster, region, resource type, or any infrastructure dimension.
GTM
  • Customer spend tracking: Provide your customers with a breakdown of their spend by user, product, time period, or any other dimension.
Product
  • Customer-of-customer tracking: Enable your customers to track spend by any dimension to support their own customers.

Finalized rated events

Finalized rated events are delivered once, after the associated invoice is finalized. They represent a stable, immutable record of event rating for that billing period. The exported rated events roll up to the finalized invoice totals for usage-based products.
DRAFT RATED EVENTS COMING SOONFinalized rated events only cover closed billing periods. We plan to add support for draft rated events, which will be delivered daily and cover all usage sent month-to-date. This will unlock near-real-time spend analytics and the ability to answer customer billing questions during an active period. Contact us via the Metronome support portal if this is important to your use case.

Schema

Joining with other tables

Caveats

  • No support for SQL billable metrics: Rated events are not generated for billable metrics defined using SQL. If your contract uses SQL-based billable metrics, those events will not appear in the rated events export.
  • Billable metrics added mid-billing period: If a billable metric is created after a billing period has already started, usage events that arrived before the metric was created will not be rated. Rated events will only be generated for usage received after the billable metric was created, and the finalized invoice will reflect only that partial coverage.
  • Voided invoices are not reflected in rated events: If an invoice is voided after it has been finalized and rated events have already been generated, the rated events for that period will not be updated or removed. To determine whether a rated event belongs to a voided invoice, join the rated_events table to the invoice table on invoice_id and filter on invoice status.
  • Commit and credit amounts may differ from finalized invoice totals: When aggregating commit or credit application amounts from rated events, the totals may not match the corresponding finalized invoice. This is because rated events use improved commit application logic that has not yet been backported to finalized invoices. This discrepancy affects only commit and credit allocation — not total billed amounts — and will be resolved when the improved logic is applied to finalized invoices in a future update.

Unrated events

Metronome processes all events delivered in the last 30 days to identify errored events, and delivers them via data export. This rolling 30-day lookback window ensures that:
  • Events received after invoice finalization are captured in error reporting.
  • Backdated usage (events timestamped before the current billing period) is identified and surfaced.
Use unrated events to detect data pipeline issues, investigate billing discrepancies, and verify that all usage has been correctly ingested and rated.

Error types

Schema

Caveats

  • No support for SQL billable metrics: Unrated events are not generated for billable metrics defined using SQL. If your contract uses SQL-based billable metrics, errors for those events will not appear in the unrated events export.