Get Started →
SOURCED FROM AWS DOCS CLIENT-SIDE ESTIMATOR NO SIGN-UP
// AMAZON CONNECT SERVICE QUOTA CHECKER & PLANNER
Verified against AWS documentation: . AWS quota defaults change infrequently — always confirm the live value for your instance in the Service Quotas console before a go-live.

Every Amazon Connect default service quota, in one place — plus a go-live estimator

Concurrent calls, chats, tasks and emails. Users, phone numbers, queues, routing profiles. Amazon Q knowledge bases, Cases, Voice ID, Customer Profiles, outbound campaigns, and API throttling. Enter your expected volumes below and this tool tells you exactly which default quotas you'll hit before your contact center goes live — so you're not filing increase requests the week of launch.

GO-LIVE ESTIMATOR

Will your expected volume hit an Amazon Connect quota?

Fill in what you expect at peak — not your average day. Everything runs in your browser; nothing is sent anywhere. Leave a field at its default (0) to skip that check.

Expected go-live parameters
Peak concurrent contacts
Default quota: 10
Default quota: 500
Default quota: 2,500
Default quota: 1,000
Org & routing resources
Default quota: 500
Default quota: 5
Default quota: 100
Default quota: 500
Default quota: 50
Default quota: 100
Default quota: 100
Default quota: 100
Default quota: 50
Default quota: 100
Conversational analytics & Voice ID
Used for analytics job & Voice ID sizing
Used for post-call analytics sizing
Outbound campaigns
Default quota: 50
Default quota: 0 — always request before launch
QuotaYou needAWS defaultStatusRecommendation
// GO-LIVE QUOTA READINESS REPORT
GO-LIVE CHECKLIST

How to plan quotas before launch

AWS's own guidance is to request quota increases as early as possible — larger requests can take up to three weeks, and an extra-large worldwide increase can take months. Work through this order.

1

Model peak, not average

Quotas like concurrent calls are consumed at your single busiest moment, not your daily average. Use your highest forecast interval, with headroom for spikes.

2

Separate account-level from resource-level

Account-level increases (e.g. instances per Region, most API throttling) apply everywhere in the account. Resource-level increases (e.g. users per instance) apply to one instance only — request each instance separately.

3

Create the instance first

You can't request a resource-level quota increase until the target Amazon Connect instance already exists. Stand up the instance early, even before flows are built.

4

Request outbound campaign capacity explicitly

Concurrent campaign active calls defaults to zero for every new account. This is the single most common launch-day surprise for outbound deployments.

5

Alert at 80% of quota

AWS's own recommendation for concurrent-email and similar quotas: configure CloudWatch alerts at 80% utilization so you request an increase before you're throttled, not after.

6

Budget lead time by size

Small increases can be approved in hours. Larger ones take up to 3 weeks. An extra-large, multi-Region increase can take months — file it as soon as your forecast is stable.

7

Re-check after go-live

Default quota values documented by AWS apply to new accounts — your account's actual applied quota can already differ. Confirm live values in the Service Quotas console rather than assuming this page's defaults still match your account.

8

Watch API throttling separately

API rate limits are shared across every user, integration, and instance in the account and Region. A reporting dashboard polling GetMetricDataV2 too aggressively can throttle your CTI or CRM integration too.

FULL REFERENCE

Complete Amazon Connect service quota reference

Every default quota documented by AWS across Amazon Connect core, AppIntegrations, Amazon Q in Connect, Cases, conversational analytics, Customer Profiles, outbound campaigns, Voice ID, and API throttling. Search or filter by category — all rows are present in the page for reference and search-engine indexing regardless of which filter is active.

Amazon Connect instance quotas

The quotas most contact centers hit first — concurrent contacts, users, phone numbers, queues, and the building blocks of routing.

Quota nameDefault valueAdjustableLevelNotes
Concurrent active calls per instance 10 Yes Resource Counts PSTN and WebRTC calls that are in a flow, queued, being handled by an agent, or an outbound call. Callbacks waiting in a callback queue and external transfers are not counted. Exceeding this returns a fast-busy (reorder) tone to the caller.
Concurrent active chats per instance 500 Yes Resource Includes SMS, WhatsApp, and Apple Messages for Business, plus chats still waiting in queue. An idle chat where the customer has gone silent still counts — use persistent chat to avoid idle chats eating into this quota.
Concurrent active tasks per instance 2,500 Yes Resource Any task not yet ended counts as active: being routed, queued, assigned to an agent, or in After Contact Work (ACW).
Concurrent active emails per instance 1,000 Yes Resource Total of emails executing through a flow, waiting in queue, assigned to agents (working or in ACW), and replies/agent-initiated emails being composed. Monitor with the ConcurrentEmails / ConcurrentEmailsPercentage CloudWatch metrics.
Users per instance 500 Yes Resource Maximum users you can create in the instance in the current Region. All 500 can be logged in and handling contacts concurrently.
Phone numbers per instance 5 Yes Resource Requires AWS CLI v2.13.20+ to view/manage via the CLI. The same quota-increase form is used to port an existing number to Amazon Connect.
Queues per instance 100 Yes Resource Standard and agent queues combined.
Maximum contacts in an agent queue 10 Yes Resource Applies per agent queue — the same cap applies to every agent queue in the instance.
Routing profiles per instance 500 Yes Resource Total routing profiles configurable in the instance.
Queue-channel combinations per routing profile 50 Yes Resource A queue counts once per channel — e.g. "Sales Voice" and "Sales Chat" are two combinations. Manually-assigned queues have their own independent 50-combination allowance.
Quick connects per instance 100 Yes Resource Phone number, queue, and user quick connects combined.
Security profiles per instance 100 Yes Resource Custom + default security profiles combined.
User hierarchy groups per instance 500 Yes Resource Total across all five hierarchy levels — there is no separate per-level cap.
Proficiencies per agent 10 Yes Resource Used for skills-based routing.
Predefined attributes per instance 150 Yes Resource Used for attribute-based routing and case fields.
Agent status per instance 50 No Not adjustable Custom + default statuses combined. Hard ceiling — plan status taxonomy carefully.
Amazon Connect instances per Region 2 Yes Account You can run multiple instances in one Region (e.g. prod + staging) — this quota, not the per-instance ones, caps how many.
AWS Lambda functions associated per instance 50 Yes Resource Functions available to invoke from a contact flow.
Flows per instance 100 Yes Resource Contact flows, flow modules are counted separately (see Modules below).
Modules per instance 200 Yes Resource Reusable flow modules referenced from contact flows.
Prompts per instance 500 Yes Resource Uploaded audio prompts.
Hours of operation per instance 100 Yes Resource
Overrides per hours of operation 50 No Not adjustable One-off date overrides (holidays, closures) per HOO record.
Associations that inherit recurring overrides per hours of operation 3 No Not adjustable
Data tables per instance 100 Yes Resource Used for dynamic, flow-referenced lookup data.
Primary attributes per data table 5 No Not adjustable
Attributes per data table 100 No Not adjustable
Values per data table 1,000 Yes Resource Rows of data per table.
Reports per instance 2,000 Yes Resource Includes every supervisor's personal saved reports — set a housekeeping policy so these don't silently pile up.
Scheduled reports per instance 100 Yes Resource
Custom metrics per instance 1,000 No Resource
Amazon Lex bots per instance 70 No Resource Classic (V1) Lex bot associations.
Amazon Lex V2 bot aliases per instance 100 Yes Resource
Task templates per instance 50 No Not adjustable
Customized fields per task template 50 No Not adjustable
Maximum duration a task can be scheduled in the future 90 days No Not adjustable
Maximum reschedules for a future-scheduled task 20 No Not adjustable
Email addresses per instance 100 (up to 500) Yes Resource Can be raised up to 500 addresses per instance.
Email addresses per email contact (To + CC) 50 total No Not adjustable Outbound emails support 1 address in To and up to 49 in CC. BCC isn't supported. Not raisable.
Email domains per instance 1 Connect domain + 100 custom No Resource
Maximum individual attachment size 20 MB (up to 100 MB) Yes Resource Configurable up to 100 MB via the admin website or API.
Active email contact expiry 14 days (up to 90) No Resource How long an email can sit active before auto-closing. Customizable up to 90 days via the connect:ContactExpiry segment attribute — not adjustable beyond that.
Active email conversation (thread) expiry 90 days No Resource A reply after 90 days starts a new thread instead of continuing the old one.

AppIntegrations quotas

All AppIntegrations quotas are account-level, applying across every Amazon Connect instance in the account and Region.

Quota nameDefault valueAdjustableLevelNotes
Data integration associations per data integration 10 Yes Account
Data integrations per Region 10 Yes Account
Event integration associations per event integration 10 Yes Account
Event integrations per Region 10 Yes Account
Third-party applications per Region 25 No Account

Amazon Q in Connect (agent assist) quotas

All Amazon Q in Connect quotas are account-level. Adjustments go through AWS Support rather than the self-service Service Quotas console.

Quota nameDefault valueAdjustableLevelNotes
Assistants 5 No Account
Knowledge bases 10 Yes Account
Assistant associations 20 No Account
Quick responses per knowledge base 1,000 Yes Account
Content items per knowledge base 5,000 Yes Account FAQs, wikis, articles, step-by-step guides.
Maximum size per document 5 MB Yes Account
Message templates per knowledge base 200 Yes Account Email, SMS, and WhatsApp templates combined.
Versions per message template 20 No Account
Attachments per email message template 10 No Account 1 MB max per attachment; 5,000,000 characters max in an email template body; 800 characters max for an SMS template.
Routing profiles assigned per quick response or template 40 No Account

Amazon Connect Cases quotas

All Cases quotas are account-level. Cases integrated with the central AWS Service Quotas console in February 2026, so utilization and increase requests are now visible directly from that console.

Quota nameDefault valueAdjustableLevelNotes
Cases domains per AWS account 5 Yes Account
Fields in a Cases domain 500 Yes Account
Options per single-select field 500 Yes Account
Layouts in a Cases domain 100 Yes Account
Templates in a Cases domain 100 Yes Account
Related items attached to a case 200 Yes Account
Files attached to a case 50 Yes Account
Case fields per case layout 100 No Account
SLAs attached to a case 10 Yes Account
Fields in a Custom-type related item 5 Yes Account

Conversational analytics (Contact Lens) quotas

All conversational analytics quotas are account-level. Post-call and chat analytics jobs run concurrently while AWS processes each finished contact — size this against your busiest hour, not your average.

Quota nameDefault valueAdjustableLevelNotes
Concurrent real-time calls with analytics 300 Yes Account Live sentiment/transcription running during the call itself.
Concurrent post-call analytics jobs 200 Yes Account A job starts after each analytics-enabled contact ends. AWS's own planning formula: (avg. call minutes × 0.4 × calls/hour) ÷ 60.
Concurrent chat analytics jobs 200 Yes Account
Concurrent automated interaction analytics jobs 20 Yes Account
Concurrent post-contact agent conversation summary jobs 10 Yes Account Shared across voice and chat.
Concurrent post-contact automated interaction summary jobs 2 Yes Account Shared across voice and chat.
Concurrent after-call agent conversation summary jobs 2 Yes Account Shared across voice and chat.
External voice analytics connectors 0 Yes Account Must be explicitly requested if you're piping in a non-Connect voice system for analytics.
Active recording sessions from external voice systems 10 Yes Account Per instance.
Evaluation questions answered with Ask AI (manual) 10 Via support ticket Account Per manually-submitted evaluation on a single contact.
Evaluation questions answered automatically (generative AI) 10 Via support ticket Account Per automated evaluation submission on a single contact.

Customer Profiles quotas

All Customer Profiles quotas are account-level and apply per supported Region.

Quota nameDefault valueAdjustableLevelNotes
Customer Profiles domains 100 Yes Account/Region
Keys per object type 10 Yes Account/Region
Maximum expiration for an object or profile 1,098 days Yes Account/Region
Calculated attributes per domain 50 No Account/Region
Event streams per domain 1 No Account/Region
Event triggers per domain 50 Yes Account/Region
Integrations per domain 50 Yes Account/Region
Object types per domain 100 Yes Account/Region
Segment snapshots per day 200 Yes Account/Region Per domain.
Total size of all objects for a profile 51,200 KB Yes Account/Region
Object / profile maximum size 250 KB No Account/Region Per single profile or profile object.
Objects per profile 1,000 Yes Account/Region
Concurrent bulk export jobs 20 No Account/Region Frees up once a job completes or fails.

Outbound campaigns quotas

All outbound campaigns quotas are account-level. Concurrent campaign active calls defaults to zero — every new account must explicitly request capacity before running its first campaign.

Quota nameDefault valueAdjustableLevelNotes
Active campaigns per instance 50 Yes Account
Total campaigns per instance 500 Yes Account Active + inactive combined.
Concurrent campaign active calls per instance 0 (must be requested) Yes Resource This is the one every new outbound deployment forgets. It defaults to zero — dials will fail with a fast-busy tone until you request a starting quota.

Voice ID quotas — end of support notice

AWS ended support for Amazon Connect Voice ID on May 20, 2026. The Voice ID console, admin website, CCP features, and API resources are no longer accessible. The figures below are kept for historical reference and for teams auditing legacy configurations only.

Quota nameDefault valueAdjustableLevelNotes
Domains 3 No Account Per account. (Feature retired May 20, 2026.)
Concurrent active sessions per domain 50 No Account AWS's scaling guidance was roughly 5% of hourly voice contact volume, e.g. 1,000 calls/hour ≈ 50 sessions, 10,000 calls/hour ≈ 500 sessions. (Feature retired May 20, 2026.)
Fraudsters per watchlist 500 No Account (Feature retired May 20, 2026.)
Watchlists per domain 3, including the default watchlist No Account (Feature retired May 20, 2026.)
Speakers per domain 100,000 No Account (Feature retired May 20, 2026.)

API throttling quotas (selected)

Amazon Connect API throttling is by AWS account and Region — not by user, not by instance. Every caller in the account shares one bucket per API. Unless listed below, every Connect API defaults to 2 requests/second sustained (RateLimit) with a burst of 5/second (BurstLimit).

Quota nameDefault valueAdjustableLevelNotes
Default (all unlisted Connect APIs) 2 TPS / burst 5 TPS Yes Account
Evaluations actions (all) 1 TPS Yes Account
GetMetricData 5 TPS / burst 8 TPS Yes Account
GetMetricDataV2 10 TPS / burst 10 TPS Yes Account
GetCurrentMetricData 5 TPS / burst 8 TPS Yes Account
SearchContacts 0.5 TPS / burst 1 TPS Yes Account The lowest default of any Connect API — a common cause of throttling in reporting integrations.
StartChatContact 5 TPS / burst 8 TPS Yes Account
StartContactStreaming / StopContactStreaming 5 TPS / burst 8 TPS Yes Account
CreateParticipant / UpdateParticipantRoleConfig 5 TPS / burst 8 TPS Yes Account
GetContactAttributes / UpdateContactAttributes / DescribeContact / StopContact / UpdateContact / ListContactReferences / BatchPutContact 10 TPS / burst 15 TPS Yes Account
TagContact / UntagContact 20 TPS / burst 25 TPS Yes Account
UpdateContactRoutingData 20 TPS / burst 20 TPS Yes Account
SendChatIntegrationEvent 17 TPS / burst 26 TPS Yes Account
CreateIntegrationAssociation / DeleteIntegrationAssociation 2 TPS / burst 5 TPS Yes Account 1 TPS for the SES_IDENTITY integration type specifically.
ListIntegrationAssociations 25 TPS / burst 50 TPS Yes Account
Participant Service: CreateParticipantConnection 6 TPS / burst 9 TPS No Instance Participant Service quotas are per-instance, not account-wide, unlike the main Connect API.
Participant Service: SendMessage / SendEvent 10 TPS / burst 15 TPS No Instance
Outbound Campaigns: CreateCampaign / StartCampaign / StopCampaign / PauseCampaign / ResumeCampaign / DeleteCampaign 1 TPS / burst 2 TPS No Account
Outbound Campaigns: PutDialRequestBatch / PutOutboundRequestBatch 10 TPS / burst 10 TPS No Account

Amazon Connect service quotas — what every architect should know before go-live

Amazon Connect enforces default resource and rate limits — "service quotas" — on almost every configurable object in an instance: concurrent contacts by channel, users, phone numbers, queues, routing profiles, flows, and integration touchpoints like Lambda and Lex. Most are adjustable through the AWS Service Quotas console with no support ticket required, but a handful — like the 50 agent statuses per instance or the 10 overrides per hours-of-operation record — are hard ceilings that only design decisions can work around. New Amazon Connect deployments most commonly get caught out by three defaults specifically: 10 concurrent active calls per instance (fine for a pilot, undersized for most production launches), 5 phone numbers per instance, and 0 concurrent campaign active calls for outbound — the last of which must be explicitly requested before a single dialer campaign can run. This page mirrors the official AWS documentation and adds a calculator so you can check your actual forecast against every default before you're troubleshooting a fast-busy tone on launch day.

What is the default concurrent call limit in Amazon Connect?
The default is 10 concurrent active calls per instance, combining PSTN and WebRTC calls that are in a flow, queued, or being handled by an agent, plus outbound calls. It's a resource-level quota, so it's adjustable per instance through the Service Quotas console — request an increase as soon as the instance exists, well before go-live.
How do I request an Amazon Connect quota increase?
Open the AWS Service Quotas console, search for Amazon Connect, select the quota you need raised, and submit a request with your desired value. Smaller increases are often approved within hours; large or multi-Region increases can take up to three weeks, or months for extra-large worldwide requests, so file early.
What's the difference between account-level and resource-level quotas?
An account-level increase applies to every Amazon Connect instance in that account and Region at once — API throttling is the most common example. A resource-level increase applies only to the specific instance you requested it for, such as users per instance or concurrent active calls, and needs to be requested separately for each instance.
How many phone numbers can I claim per Amazon Connect instance?
The default is 5 phone numbers per instance, and it's an adjustable resource-level quota. AWS CLI version 2.13.20 or higher is required to view and manage this quota from the command line, and the same request form is used to port an existing number to Amazon Connect.
What's the maximum number of users per Amazon Connect instance?
The default is 500 users per instance, and all 500 can be logged in and handling contacts concurrently. It's adjustable — organizations with larger contact centers routinely request and receive higher limits.
Why does my outbound campaign fail with a fast-busy tone immediately after launch?
Concurrent campaign active calls per instance defaults to zero for every new account. Dials will fail with a fast-busy tone until you explicitly request a starting quota for this specific limit — it's easy to miss because every other campaign-related quota (active campaigns, total campaigns) already has a usable default.
Which Amazon Connect quotas cannot be increased?
A handful are hard ceilings: 50 agent statuses per instance, 50 overrides per hours-of-operation record, 5 primary attributes and 100 attributes per data table, 50 task templates per instance, and the email address limits per message (50 total across To/CC, no BCC). These require a design workaround rather than a quota request.
How is the concurrent active calls quota actually counted?
Calls being handled by a flow, waiting in queue, being handled by an agent, and outbound calls all count. Callbacks still waiting in a callback queue don't count until offered to an agent, and external transfers aren't counted either. You can estimate your currently configured quota using the ConcurrentCallsPercentage CloudWatch metric.
How many Amazon Connect instances can I run in one Region?
The default is 2 instances per Region per account, and it's an account-level, adjustable quota — commonly raised to support separate production and staging/UAT instances, or multiple business units in the same Region.
How do I estimate concurrent post-call analytics jobs for Contact Lens?
AWS's planning formula is (average call duration in minutes × 0.4 × calls per hour) ÷ 60, since analysis of a finished call typically takes about 40% of its length. The default quota is 200 concurrent jobs — high-volume centers doing 3,000+ calls/hour at 10-minute average duration will land close to that default and should request an increase ahead of time.
Is Amazon Connect Voice ID still available?
No — AWS ended support for Amazon Connect Voice ID on May 20, 2026. The console, admin website, CCP integration, and API resources are no longer accessible. The historical default quotas (3 domains, 50 concurrent sessions per domain, 100,000 speakers per domain) are preserved on this page for teams auditing legacy configurations only.
Are Amazon Connect API rate limits shared across my integrations?
Yes — API throttling quotas are by AWS account and Region, not by user or instance. Every user, every instance, and every integration calling the same API in that account and Region shares one throttle bucket, so a reporting job and a CTI integration calling GetMetricDataV2 at the same time can throttle each other.

Found this tool helpful?

Share it with your team!

Twitter Facebook LinkedIn