For the complete documentation index, see llms.txt. This page is also available as Markdown.

Rate Limits & Throughput

How ComboBuilder handles request volume, and what to expect at peak traffic

The ComboBuilder API is built for high request volumes, including the traffic peaks that come with major tournaments. This page explains how request volume is counted, the capacity provisioned for your account, and how the API behaves when traffic runs high.

How request volume is counted

Request volume is measured across the whole ComboBuilder API, not per endpoint. Calls to /v1/markets, /v1/combobuilder, and /v1/combobuilder/auto all count toward the same combined total. There is no separate allowance for each endpoint.

Most of your volume comes from /v1/combobuilder, because a bettor's bet slip is repriced every time they add or change a selection.

Provisioned capacity

Each operator is provisioned with a throughput baseline of around 2,000 requests per second (RPS), shared across all ComboBuilder endpoints. This is the capacity allocated to you by default, not a fixed ceiling.

If you expect sustained traffic above this baseline — for example around a major tournament — share your projected peak RPS with your CSM in advance. We provision additional capacity ahead of the event so your traffic is covered.

Tell your CSM your expected peak volume before a high-traffic event. Advance notice lets us allocate capacity in time.

What happens at high volume

Under normal operation, ComboBuilder does not hard-block your traffic with a rate-limit error. The API scales to absorb spikes in demand.

If sustained traffic exceeds the capacity provisioned for you, requests are not rejected with a dedicated rate-limit response. Instead, response times rise, and any request that cannot complete in time returns a standard server response:

Response
Meaning
What to do

503 Service Unavailable

The service is momentarily saturated

Retry with exponential backoff

504 Gateway Timeout

The request did not complete in time

Retry with exponential backoff

Build standard resilience into your integration: retry on 503 and 504 with exponential backoff, and treat both as transient. If you see them consistently, contact your CSM to review your provisioned capacity.

Last updated

Was this helpful?