Skip to content

Usage and limits

Find what drives credit consumption and plan capacity as your application grows.

Budget for growth and explain unexpected spending by finding the workloads that use your credits. All services in your organization share one credit balance, and usage shows which product, API key, or resource did the work.

How usage is recorded

Each service records the work done for your organization, attributed to the API key that sent it or the resource that ran. Records settle into hourly UTC totals, which can take a few minutes; settlement draws credits from your organization's balance and updates Usage and each key's Last used date.

Because usage is attributed to the API key that sent the work, a separate key per workload is what lets you tell workloads apart later.

Find what is driving usage

  1. Open Usage and choose the time range you want to investigate.
  2. Compare daily totals for trends, or hourly totals to narrow down a spike.
  3. Filter by product and API key to isolate an application.
  4. For indexing costs, filter by resource to find the Subgraph responsible.
  5. Use event-type filters to separate different kinds of work within a product.

After rotating a key, use the revocation timestamp in the API key dropdown to distinguish the old key from its replacement. Revoked keys show their name followed by (Revoked …), with the exact date, time, and your time zone, and remain selectable so you can inspect their historical usage.

Read usage times

Usage may take a few minutes to appear after work completes. The current hour and day are still in progress, so their totals can change. Use completed periods when comparing consumption across days.

Hourly and daily totals follow UTC boundaries. Times from the past week show relative to now, such as 2 hours ago, and older ones show the date. Hover over, focus, or tap a time to see its full date, time, and time zone.

If activity is missing, check the date range and filters. If it still does not appear, contact GraphOps support with your organization, product, API key label, approximate UTC time, and request IDs. Include your local time zone too.

Control spending

  • Estimate a workload with the service's current rates before running it at scale.
  • Cap retries and avoid repeating completed requests. Successful retries can count as additional work.
  • Split large backfills into ranges you can pause, and check usage after the first range.
  • Set low-credit alerts so the right people hear before the balance runs out.

Plan for more traffic

Limits apply across all callers in your organization, not per API key. Check your service's usage page and plan before increasing traffic, concurrency, or backfill size.

QuestionWhere to find the answer
How many credits remain?Billing in the platform.
Which workloads use them?Usage in the platform.
What does each service charge?The service usage pages in the next steps below.
How much traffic can I send?Your service limits and plan. Contact support for a capacity increase.

Next steps