specification: API Commons Rate Limits
specificationVersion: '0.1'
schema: https://raw.githubusercontent.com/api-evangelist/interface-research/main/schema/api-commons.yml#/$defs/RateLimits
provider: Amazon Connect
providerId: amazon-connect
created: '2026-05-04'
# Provenance stamped 2026-08-11: this artifact was written by the API Evangelist
# bulk sweep dated 2026-05-04, not harvested from the provider. See roadmap#35.
method: generated
modified: '2026-05-05'
reconciled: true
tags:
  - Cloud
  - Contact Center
  - Customer Engagement
  - AWS
  - Rate Limiting
description: Amazon Connect API throttling is enforced per AWS account and per Region, not per user or
  instance. The default Connect Service API rate is 2 RPS with a burst of 5 RPS, with per-API exceptions
  documented (metric, contact, integration, and search APIs have higher limits). Contact Lens, Cases,
  Customer Profiles, Voice ID, AI Agents, Outbound Campaigns, and Participant APIs each publish their
  own throttling tables. Most quotas are adjustable via AWS Service Quotas requests; small increases approve
  in hours, larger increases can take up to three weeks. Resource-level quotas (users, queues, concurrent
  calls/chats) are per-instance.
sources:
  - https://docs.aws.amazon.com/connect/latest/adminguide/amazon-connect-service-limits.html
responseCodes:
  throttled: 429
  throttledLegacy: 503
limits:
  - name: Connect Service API (default)
    scope: account/region
    metric: requests_per_second
    limit: 2
    timeFrame: second
    notes: Default RateLimit for most operations; BurstLimit 5. Exceptions documented per API.
  - name: GetMetricData
    scope: account/region
    metric: requests_per_second
    limit: 5
    timeFrame: second
    notes: BurstLimit 8.
  - name: GetMetricDataV2
    scope: account/region
    metric: requests_per_second
    limit: 10
    timeFrame: second
    notes: BurstLimit 10.
  - name: GetCurrentMetricData
    scope: account/region
    metric: requests_per_second
    limit: 5
    timeFrame: second
    notes: BurstLimit 8.
  - name: SearchContacts
    scope: account/region
    metric: requests_per_second
    limit: 0.5
    timeFrame: second
    notes: BurstLimit 1.
  - name: Contact attribute / lifecycle APIs
    scope: account/region
    metric: requests_per_second
    limit: 10
    timeFrame: second
    notes: GetContactAttributes, UpdateContactAttributes, DescribeContact, StopContact, UpdateContact,
      ListContactReferences, BatchPutContact - BurstLimit 15.
  - name: Tag / Untag / UpdateContactRoutingData
    scope: account/region
    metric: requests_per_second
    limit: 20
    timeFrame: second
    notes: TagContact / UntagContact BurstLimit 25; UpdateContactRoutingData BurstLimit 20.
  - name: Evaluations actions
    scope: account/region
    metric: requests_per_second
    limit: 1
    timeFrame: second
  - name: Connect AI Agents (default)
    scope: account
    metric: requests_per_second
    limit: 10
    timeFrame: second
    notes: 10 TPS default with per-API exceptions.
  - name: Customer Profiles (write)
    scope: account
    metric: requests_per_second
    limit: 100
    timeFrame: second
    notes: SearchProfiles, ListProfileObjects, CreateProfile, UpdateProfile, MergeProfiles, etc. = 100
      TPS. Domain CRUD APIs = 1-5 TPS.
  - name: Concurrent active calls per instance
    scope: instance
    metric: concurrent_requests
    limit: 10
    notes: PSTN + WebRTC; adjustable.
  - name: Concurrent active chats per instance
    scope: instance
    metric: concurrent_requests
    limit: 500
    notes: Includes SMS, WhatsApp, Apple Messages for Business; idle chats count - use persistent chat
      to avoid this.
  - name: Concurrent active emails per instance
    scope: instance
    metric: concurrent_requests
    limit: 1000
    notes: Adjustable.
  - name: Concurrent active tasks per instance
    scope: instance
    metric: concurrent_requests
    limit: 2500
    notes: Adjustable.
  - name: Users per instance
    scope: instance
    metric: count
    limit: 500
    notes: Adjustable.
  - name: Phone numbers per instance
    scope: instance
    metric: count
    limit: 5
    notes: Adjustable.
  - name: Connect Cases - GetCase
    scope: account
    metric: requests_per_second
    limit: 4
    timeFrame: second
    notes: BurstLimit 10.
  - name: Connect Cases - default
    scope: account
    metric: requests_per_second
    limit: 2
    timeFrame: second
    notes: BurstLimit 10 for create/list APIs; 2 for update.
  - name: Voice ID - EvaluateSession
    scope: account
    metric: requests_per_second
    limit: 60
    timeFrame: second
  - name: Outbound Campaigns - default
    scope: account/region
    metric: requests_per_second
    limit: 2
    timeFrame: second
    notes: BurstLimit 5; campaign-mutation APIs limited to 1 RPS / Burst 2; DescribeCampaign 25/35.
  - name: Participant Service - SendEvent / SendMessage
    scope: instance
    metric: requests_per_second
    limit: 10
    timeFrame: second
    notes: BurstLimit 15.
  - name: Contact Lens - ListRealtimeContactAnalysisSegmentsV2
    scope: account
    metric: requests_per_second
    limit: 2
    timeFrame: second
    notes: BurstLimit 5.
policies:
  - name: Per-Account/Region Throttle Bucket
    description: Throttling is shared across all users and instances inside the same AWS account and Region;
      requests across instances share the same bucket.
  - name: Service Quotas Adjustment
    description: Most quotas are adjustable via AWS Service Quotas. Small increases approve in hours;
      larger or worldwide increases can take up to three weeks. Open a quota-increase request before going
      live.
  - name: Resource vs Account Quotas
    description: Resource-level quotas (users, queues, concurrent calls/chats/emails) apply per Connect
      instance. Account-level quotas (API TPS, instance count) apply across instances in a Region.
  - name: Backoff Strategy
    description: Implement exponential backoff with jitter on 429/503 throttling responses; AWS SDKs handle
      this by default.
  - name: Plan Ahead
    description: Configure CloudWatch alerts at 80% of quota limits; coordinate with AWS support before
      production go-live for capacity planning.
maintainers:
  - FN: Kin Lane
    email: kin@apievangelist.com