Usage Explorer

Usage Explorer shows telemetry about your own product — API requests, orders, active seats, GB processed, or any other count you send in — separately from cost data. Cost sources tell you what you spent; usage tells you how much work that spend produced.

Where the data comes from

Usage data reaches Plutus one of three ways:

• Pushed from your own backend. You call the usage ingestion API with an API key and send events as they happen (or on whatever schedule you choose). See Usage ingestion API for the request format and authentication.
• Pulled from a metered-billing provider. If you connect a usage source like Stripe Billing usage records, Metronome, or Chargebee, Plutus syncs usage on that source’s own schedule, the same way a cost source syncs.
• Emitted as a byproduct of a cost or event source you’ve already connected. A number of cost and event sources — including AWS, GCP, OpenAI, Anthropic, GitHub, GitLab, Vercel, and others — write usage metrics (token counts, compute hours, deploy counts, seat counts, and similar) straight from the same API call their normal cost or event sync already makes. There’s nothing extra to connect: if a source’s page lists usage metrics, they start showing up here as soon as that source is syncing. See each source’s own page under Sources & integrations for whether it emits usage metrics and which ones.

In every case, each data point has a metric name, a numeric quantity, and a timestamp. You can send more than one metric — for example, both api_requests and orders_completed — and choose which one to look at in the explorer.

Browsing usage

Usage Explorer works like Cost Explorer: pick a date range and narrow with Filters and Group by, the same two controls that page uses, adapted to usage’s simpler shape — one grain (metric) rather than several dimensions to pivot between.

• Filters narrows which sources feed the chart, and (once a source is broken out — see below) which of its individual metrics are included.
• Group by breaks a source into one row per metric instead of one lumped total. A source you haven’t broken out still shows up, just as a single combined line.

As with Cost Explorer, narrowing and grouping are independent — breaking a source out doesn’t change which sources are in scope, and filtering doesn’t change how a broken-out source is grouped.

There’s no cost figure on this page by itself — usage on its own is just a count. To turn a count into a cost figure, use unit economics.

Feeding unit economics

Unit economics divides a cost figure by a usage metric to produce a ratio, such as cost per API request or infra cost per customer. It’s a separate page (Unit economics), but the usage side of that ratio is exactly the data you see here — a unit metric can only use a metric that’s already showing up in Usage Explorer.

If a metric you expect to use for unit economics isn’t showing data, check that the usage source or ingestion integration sending it is connected and has synced recently before troubleshooting the unit metric itself.