From 5e60b19bc45ca52b084be053e01ce1e4a92b6041 Mon Sep 17 00:00:00 2001 From: Scout <3745266+relay-scout[bot]@users.noreply.github.com> Date: Mon, 17 Aug 2026 20:57:37 +0000 Subject: [PATCH 1/2] docs: explain API rate-limit upgrades (INT2-1253) Document the standard production upgrade, required capacity intake, and whole-key scope so integrators know what to include in a review request. Note that requests above the elevated limits need additional review. Linear: INT2-1253 --- references/api/api-keys.mdx | 4 ++-- references/api/api_core_concepts/handling-rate-limits.mdx | 4 ++-- 2 files changed, 4 insertions(+), 4 deletions(-) diff --git a/references/api/api-keys.mdx b/references/api/api-keys.mdx index 8c183c6..e008f97 100644 --- a/references/api/api-keys.mdx +++ b/references/api/api-keys.mdx @@ -42,7 +42,7 @@ The following limits apply per API key: ## Elevated Rate Limits -Higher limits are available on request, applied per key: +Higher limits are available on request. The standard production upgrade raises a key from the default 50 requests per minute to the elevated limits below. Relay applies an upgrade to the complete API key, not to an individual endpoint such as `/quote`. | Endpoint | Limit | | ---------------------- | ----------------------- | @@ -54,7 +54,7 @@ Higher limits are available on request, applied per key: `/requests` and `/requests/v2` share the 10 rps bucket; `/requests/v3` has its own 20 rps bucket. -To request higher limits, get in touch through the support widget in the [Relay Dashboard](https://dashboard.relay.link). +To request this upgrade, get in touch through the support widget in the [Relay Dashboard](https://dashboard.relay.link). Include your production use case, why the current limit is insufficient, your expected average and peak request rates, and the API endpoints you plan to use. Requests above the elevated limits require additional review. Already hitting a limit? See [Handling Rate Limits](/references/api/api_core_concepts/handling-rate-limits) for how to recover from `429` responses and reduce your request volume before requesting an upgrade. diff --git a/references/api/api_core_concepts/handling-rate-limits.mdx b/references/api/api_core_concepts/handling-rate-limits.mdx index 1c2b705..0d9def2 100644 --- a/references/api/api_core_concepts/handling-rate-limits.mdx +++ b/references/api/api_core_concepts/handling-rate-limits.mdx @@ -37,6 +37,6 @@ Either option eliminates polling for status. Keep a low-frequency reconciliation ## Request an elevated rate limit -If your workflow is already optimized and you've moved off polling, reach out to our support team to request an elevated rate limit. Elevated limits are applied per key. The [Elevated Rate Limits](/references/api/api-keys#elevated-rate-limits) table lists what's available. +If your workflow is already optimized and you've moved off polling, reach out to our support team to request an elevated rate limit. The standard production upgrade raises a key from 50 requests per minute to 10 requests per second for `/quote`; the [Elevated Rate Limits](/references/api/api-keys#elevated-rate-limits) table lists the limits for other endpoints. Relay applies an upgrade to the complete API key, not to an individual endpoint. -Get in touch through the support widget in the [Relay Dashboard](https://dashboard.relay.link). Tell us the label of the API key as it appears in the Dashboard rather than the key itself, the endpoints involved, your expected sustained and peak volume in requests per second, and what you've already optimized. +Get in touch through the support widget in the [Relay Dashboard](https://dashboard.relay.link). Tell us the label of the API key as it appears in the Dashboard rather than the key itself, your production use case, why the current limit is insufficient, your expected average and peak request rates, and every endpoint you plan to use. Requests above the elevated limits require additional review. From 2efd9377a0ba7fb37217f1d620e390b321e6e62e Mon Sep 17 00:00:00 2001 From: Scout <3745266+relay-scout[bot]@users.noreply.github.com> Date: Mon, 17 Aug 2026 21:01:12 +0000 Subject: [PATCH 2/2] docs: document 429 polling backoff (INT2-1253) Narrow the rate-limit guidance to the retry behavior clients need when polling receives a 429 response. Linear: INT2-1253 Co-authored-by: Pedro Cunha <6723339+pedromcunha@users.noreply.github.com> --- references/api/api-keys.mdx | 4 ++-- references/api/api_core_concepts/handling-rate-limits.mdx | 6 +++--- 2 files changed, 5 insertions(+), 5 deletions(-) diff --git a/references/api/api-keys.mdx b/references/api/api-keys.mdx index e008f97..8c183c6 100644 --- a/references/api/api-keys.mdx +++ b/references/api/api-keys.mdx @@ -42,7 +42,7 @@ The following limits apply per API key: ## Elevated Rate Limits -Higher limits are available on request. The standard production upgrade raises a key from the default 50 requests per minute to the elevated limits below. Relay applies an upgrade to the complete API key, not to an individual endpoint such as `/quote`. +Higher limits are available on request, applied per key: | Endpoint | Limit | | ---------------------- | ----------------------- | @@ -54,7 +54,7 @@ Higher limits are available on request. The standard production upgrade raises a `/requests` and `/requests/v2` share the 10 rps bucket; `/requests/v3` has its own 20 rps bucket. -To request this upgrade, get in touch through the support widget in the [Relay Dashboard](https://dashboard.relay.link). Include your production use case, why the current limit is insufficient, your expected average and peak request rates, and the API endpoints you plan to use. Requests above the elevated limits require additional review. +To request higher limits, get in touch through the support widget in the [Relay Dashboard](https://dashboard.relay.link). Already hitting a limit? See [Handling Rate Limits](/references/api/api_core_concepts/handling-rate-limits) for how to recover from `429` responses and reduce your request volume before requesting an upgrade. diff --git a/references/api/api_core_concepts/handling-rate-limits.mdx b/references/api/api_core_concepts/handling-rate-limits.mdx index 0d9def2..6eab693 100644 --- a/references/api/api_core_concepts/handling-rate-limits.mdx +++ b/references/api/api_core_concepts/handling-rate-limits.mdx @@ -19,7 +19,7 @@ Quote traffic is the next place to look. `/quote` carries the tightest default l Validation belongs on the client too. Amounts below Relay's minimum come back as [`AMOUNT_TOO_LOW`](/references/api/api_core_concepts/handling-errors), so a sub-cent value a user types should fail your own check before it turns into a quote call. Where your interface already knows the user's balance, apply the same treatment to amounts they can't cover. -If you poll for status, widen the interval and stop polling once a request reaches a terminal status (`success`, `failure`, or `refund`). When you need many records at once, [`GET /requests/v3`](/references/api/get-requests) retrieves them in a single filtered call instead of one call per request ID. +If a polling request returns a `429` response, retry with exponential backoff instead of retrying immediately. --- @@ -37,6 +37,6 @@ Either option eliminates polling for status. Keep a low-frequency reconciliation ## Request an elevated rate limit -If your workflow is already optimized and you've moved off polling, reach out to our support team to request an elevated rate limit. The standard production upgrade raises a key from 50 requests per minute to 10 requests per second for `/quote`; the [Elevated Rate Limits](/references/api/api-keys#elevated-rate-limits) table lists the limits for other endpoints. Relay applies an upgrade to the complete API key, not to an individual endpoint. +If your workflow is already optimized and you've moved off polling, reach out to our support team to request an elevated rate limit. Elevated limits are applied per key. The [Elevated Rate Limits](/references/api/api-keys#elevated-rate-limits) table lists what's available. -Get in touch through the support widget in the [Relay Dashboard](https://dashboard.relay.link). Tell us the label of the API key as it appears in the Dashboard rather than the key itself, your production use case, why the current limit is insufficient, your expected average and peak request rates, and every endpoint you plan to use. Requests above the elevated limits require additional review. +Get in touch through the support widget in the [Relay Dashboard](https://dashboard.relay.link). Tell us the label of the API key as it appears in the Dashboard rather than the key itself, the endpoints involved, your expected sustained and peak volume in requests per second, and what you've already optimized.