Amazon API Gateway limits and quotas
By default API Gateway allows 10,000 requests per second per account per Region, shared across REST, HTTP and WebSocket APIs, with a burst bucket of 5,000 requests; some newer Regions start at 2,500 RPS. Request and response payloads are capped at 10 MB, REST APIs allow 300 resources and 10 stages per API, and integrations time out at 29 seconds (30 for HTTP APIs). The throttle quota is adjustable; payload and most size limits are not.
Account-level throttling
| Quota | Default | Adjustable |
|---|---|---|
| Throttle per account per Region (REST, HTTP, WebSocket combined) | 10,000 RPS | Yes |
| Burst (token bucket capacity) | 5,000 requests | No, set by AWS from the RPS quota |
| Throttle in newer Regions (for example Cape Town, Milan, Jakarta, Zurich, Spain) | 2,500 RPS, burst 1,250 | Yes |
This is an account-wide pool. One busy API can throttle every other API in the same account and Region. Stage and method throttling, and usage plans, can only lower limits within that pool, never raise them. When the pool is empty clients receive 429 Too Many Requests with {"message":"Too Many Requests"}; the 429 page separates account throttling from usage-plan and Lambda throttling.
REST API quotas worth knowing
| Quota | Default | Adjustable |
|---|---|---|
| Payload size | 10 MB | No |
| Integration timeout | 50 ms to 29 s | Yes for Regional and private APIs |
| Resources per API | 300 | Yes |
| Stages per API | 10 | Yes |
| Regional APIs per Region | 600 | No |
| Edge-optimized APIs per Region | 120 | No |
| Private APIs per account per Region | 600 | No |
| Total size of all header values | 20,480 bytes (8,000 for private APIs) | No |
| Mapping template size | 300 KB | No |
| Cached response size | 1,048,576 bytes | No |
| API caching TTL | 300 s default, 0 to 3,600 | No (upper bound) |
| Usage plans per account per Region | 300 | Yes |
| VPC links per account per Region | 20 | Yes |
| Idle connection timeout | 310 seconds | No |
Control-plane calls are throttled separately and cannot be raised: for example CreateDeployment is limited to one request every 5 seconds per account, which matters for pipelines that deploy many APIs in parallel.
HTTP API differences
HTTP APIs share the account throttle but have their own limits: 300 routes and 300 integrations per API, a fixed 30-second maximum integration timeout, a 10 MB payload, 10,240 bytes for the request line plus headers, 10 stages and 10 authorizers per API. Lambda authorizers must answer within 10 seconds, and JWKS endpoints within 1,500 ms.
Check usage and raise the throttle
# Requests per second for one API over the last hour (1-minute sums / 60)
aws cloudwatch get-metric-data --start-time 2026-11-17T08:00:00Z --end-time 2026-11-17T09:00:00Z \
--metric-data-queries '[{"Id":"c","MetricStat":{"Metric":{"Namespace":"AWS/ApiGateway","MetricName":"Count","Dimensions":[{"Name":"ApiName","Value":"orders-api"}]},"Period":60,"Stat":"Sum"}}]'
# Find the throttle quota and request more
aws service-quotas list-service-quotas --service-code apigateway \
--query "Quotas[?contains(QuotaName,'Throttle')].[QuotaName,QuotaCode,Value]" --output tableAdd up Count across all APIs in the Region when judging headroom, since the quota is shared. An API Gateway Count alarm on the busiest API is a cheap early warning.
Designing around the limits
- Split very high-traffic APIs into separate accounts if one service should never be able to starve another.
- Use per-method throttling and usage plans to protect backends and to stop one client using the whole pool.
- Enable caching for read-heavy GET methods to take load off the backend and its own limits.
- Send large uploads and downloads straight to S3 with presigned URLs instead of through the 10 MB payload limit.
- For long work, return 202 and process asynchronously rather than waiting on the 29-second timeout.
Questions
What is the API Gateway rate limit?
10,000 requests per second per account per Region by default, with a burst capacity of 5,000, shared across all REST, HTTP and WebSocket APIs. Some newer Regions default to 2,500.
What is the maximum payload size for API Gateway?
10 MB for REST and HTTP APIs. Lambda proxy integrations are further limited by Lambda's 6 MB synchronous payload.
How do I increase the API Gateway throttling limit?
Request an increase to the account-level throttle quota in Service Quotas for Amazon API Gateway in the Region concerned.
Throttling at API Gateway shows up to users as 429s long before anyone looks at a dashboard. CloudConsole, a phone app for AWS, is being built to put API error alarms and request counts on your phone.
About the app → · Download on the App Store