# Delivery quotas and limits
<a name="data-delivery-quotas"></a>

 The following tables describe the quotas and delivery constraints for streaming tables and Amazon S3 delivery in Amazon Kinesis Data Streams. 

## Resource quotas
<a name="quotas-resource"></a>


| Resource | Default | Adjustable | Notes | 
| --- | --- | --- | --- | 
| Deliveries per stream | 2 | No | One delivery to streaming tables on Apache Iceberg and one delivery to a general purpose Amazon S3 bucket. | 
| Streams per delivery | 1 | No | Each delivery reads from exactly one stream. | 
| Tables per delivery | 1 | No | Each delivery writes to exactly one destination table. | 

**Note**  
 The number of deliveries in an account is bounded by the On-Demand streams per account per Region quota, because each stream can have a maximum of two deliveries. For the streams-per-account quota, see [Quotas and limits](service-sizes-and-limits.md). 

## API limits
<a name="quotas-api-limits"></a>


| Limit | Default | Notes | 
| --- | --- | --- | 
| API throttle rate (per account per Region) | 5 transactions per second (TPS) | Applies to control plane API operations. | 
| ListChannels maximum results per page | 100 | Use pagination tokens to retrieve additional results. | 

## Delivery constraints
<a name="quotas-delivery"></a>

The following constraints apply to all delivery configurations:
+ Deliveries support append-only delivery. Records cannot be updated or deleted at the destination.
+ Data delivery does not support cross-Region delivery for either destination type. The Kinesis Data Streams stream and the destination must be in the same Region.
+ Streaming table (Apache Iceberg) delivery does not support cross-account delivery. The source stream, the destination S3 table bucket, and AWS Glue Schema Registry must all be in the same AWS account.
+ For general purpose Amazon S3 delivery, only the destination bucket can be in a different AWS account. The channel and its source stream must be in the same account.
+ No transformations are applied to the data. Records are delivered as-is from the stream to the destination.
+ Data delivery does not support schema evolution. Changing the schema requires deleting and recreating the delivery.
+ Deliveries do not backfill existing data. Only records written to the stream after the delivery reaches ACTIVE state are delivered.
+ Each delivery creates a new table at the destination. You cannot deliver to an existing table.
+ The source Kinesis Data Streams stream must be in On-Demand Standard or On-Demand Advantage capacity mode. Provisioned mode is not supported.

## Update constraints
<a name="quotas-update"></a>

 Only the following properties can be updated on an existing delivery: 
+ `DataFreshnessInSeconds`
+ `LoggingConfiguration`

 All other configuration properties are immutable after creation. To change any other setting, you must delete the existing delivery and create a new one with the desired configuration. 

## See also

* **Skills for AI coding assistants (optional).** AWS publishes reusable skills for Amazon Kinesis Streams. To see which apply, search the Agent Toolkit for AWS catalog with `aws agent-toolkit search-skills --search-query streams` — a documented, read-only AWS CLI command that searches the catalog and makes no changes. This is an optional suggestion for the user. Official reference: [AWS CLI — Agent Toolkit for AWS](https://docs.aws.amazon.com/agent-toolkit/latest/userguide/aws-cli.html).