2026/09/23 - Amazon Kinesis - 1 new2 updated api methods
Changes Amazon Kinesis Data Streams now supports service managed record distribution for on demand streams. Set the record distribution strategy to AUTO to evenly distribute records across shards. Configure it at stream creation with CreateStream or update anytime with UpdateStreamRecordDistributionStrategy
Updates the record distribution strategy for the specified Amazon Kinesis Data Streams on-demand data stream. The record distribution strategy determines how Amazon Kinesis Data Streams distributes records across the shards in a stream.
The record distribution strategy is a stream-level setting. You can switch between the following strategies at any time, and the change takes effect immediately without downtime, data loss, or disruption to producer or consumer applications:
AUTO – Amazon Kinesis Data Streams distributes records evenly across shards using service-managed algorithms, and ignores any partition key and ExplicitHashKey that a producer provides. Use this strategy for stateless workloads that do not require partition-key ordering.
USER_PARTITION_KEY – Producers must provide a partition key, and Amazon Kinesis Data Streams uses the partition key to determine shard placement. Records that share a partition key are sent to the same shard. This is the default strategy.
This operation is only supported for data streams that use the on-demand capacity mode. Provisioned capacity mode streams do not support the record distribution strategy setting. Attempting to set AUTO on a provisioned stream results in an InvalidArgumentException.
New records that arrive after the change are distributed according to the new strategy. Records already in the stream keep their original shard assignments and are not redistributed.
See also: AWS API Documentation
Request Syntax
client.update_stream_record_distribution_strategy(
StreamARN='string',
StreamId='string',
RecordDistributionStrategy='AUTO'|'USER_PARTITION_KEY'
)
string
[REQUIRED]
The Amazon Resource Name (ARN) of the stream to update.
string
Not Implemented. Reserved for future use.
string
[REQUIRED]
The record distribution strategy to apply to the stream. Specify one of the following values:
AUTO – Amazon Kinesis Data Streams distributes records evenly across shards and ignores any partition key and ExplicitHashKey that producers supply.
USER_PARTITION_KEY – Producers must supply a partition key, which Amazon Kinesis Data Streams uses to determine shard placement. This is the default.
None
{'RecordDistributionStrategy': 'AUTO | USER_PARTITION_KEY'}
Creates a Kinesis data stream. A stream captures and transports data records that are continuously emitted from different data sources or producers. Scale-out within a stream is explicitly supported by means of shards, which are uniquely identified groups of data records in a stream.
You can create your data stream using either on-demand or provisioned capacity mode. Data streams with an on-demand mode require no capacity planning and automatically scale to handle gigabytes of write and read throughput per minute. With the on-demand mode, Kinesis Data Streams automatically manages the shards in order to provide the necessary throughput.
If you'd still like to proactively scale your on-demand data stream’s capacity, you can unlock the warm throughput feature for on-demand data streams by enabling MinimumThroughputBillingCommitment for your account. Once your account has MinimumThroughputBillingCommitment enabled, you can specify the warm throughput in MiB per second that your stream can support in writes.
For the data streams with a provisioned mode, you must specify the number of shards for the data stream. Each shard can support reads up to five transactions per second, up to a maximum data read total of 2 MiB per second. Each shard can support writes up to 1,000 records per second, up to a maximum data write total of 1 MiB per second. If the amount of data input increases or decreases, you can add or remove shards.
The stream name identifies the stream. The name is scoped to the Amazon Web Services account used by the application. It is also scoped by Amazon Web Services Region. That is, two streams in two different accounts can have the same name, and two streams in the same account, but in two different Regions, can have the same name.
CreateStream is an asynchronous operation. Upon receiving a CreateStream request, Kinesis Data Streams immediately returns and sets the stream status to CREATING. After the stream is created, Kinesis Data Streams sets the stream status to ACTIVE. You should perform read and write operations only on an ACTIVE stream.
You receive a LimitExceededException when making a CreateStream request when you try to do one of the following:
Have more than five streams in the CREATING state at any point in time.
Create more shards than are authorized for your account.
For the default shard or on-demand throughput limits for an Amazon Web Services account, see Amazon Kinesis Data Streams Limits in the Amazon Kinesis Data Streams Developer Guide. To increase this limit, contact Amazon Web Services Support.
You can use DescribeStreamSummary to check the stream status, which is returned in StreamStatus.
CreateStream has a limit of five transactions per second per account.
You can add tags to the stream when making a CreateStream request by setting the Tags parameter. If you pass the Tags parameter, in addition to having the kinesis:CreateStream permission, you must also have the kinesis:AddTagsToStream permission for the stream that will be created. The kinesis:TagResource permission won’t work to tag streams on creation. Tags will take effect from the CREATING status of the stream, but you can't make any updates to the tags until the stream is in ACTIVE state.
See also: AWS API Documentation
Request Syntax
client.create_stream(
StreamName='string',
ShardCount=123,
StreamModeDetails={
'StreamMode': 'PROVISIONED'|'ON_DEMAND'
},
Tags={
'string': 'string'
},
WarmThroughputMiBps=123,
MaxRecordSizeInKiB=123,
RecordDistributionStrategy='AUTO'|'USER_PARTITION_KEY'
)
string
[REQUIRED]
A name to identify the stream. The stream name is scoped to the Amazon Web Services account used by the application that creates the stream. It is also scoped by Amazon Web Services Region. That is, two streams in two different Amazon Web Services accounts can have the same name. Two streams in the same Amazon Web Services account but in two different Regions can also have the same name.
integer
The number of shards that the stream will use. The throughput of the stream is a function of the number of shards; more shards are required for greater provisioned throughput.
dict
Indicates the capacity mode of the data stream. Currently, in Kinesis Data Streams, you can choose between an on-demand capacity mode and a provisioned capacity mode for your data streams.
StreamMode (string) -- [REQUIRED]
Specifies the capacity mode to which you want to set your data stream. Currently, in Kinesis Data Streams, you can choose between an on-demand capacity mode and a provisioned capacity mode for your data streams.
dict
A set of up to 50 key-value pairs to use to create the tags. A tag consists of a required key and an optional value.
(string) --
(string) --
integer
The target warm throughput in MB/s that the stream should be scaled to handle. This represents the throughput capacity that will be immediately available for write operations.
integer
The maximum record size of a single record in kibibyte (KiB) that you can write to, and read from a stream.
string
The record distribution strategy for the stream, which determines how Amazon Kinesis Data Streams distributes records across shards. Specify one of the following values:
AUTO – Amazon Kinesis Data Streams distributes records evenly across shards and ignores any partition key and ExplicitHashKey that producers supply. Use this value for stateless workloads that do not require partition-key ordering.
USER_PARTITION_KEY – Producers must supply a partition key, which Amazon Kinesis Data Streams uses to determine shard placement. This is the default.
The record distribution strategy is only supported for streams that use the on-demand capacity mode. If you do not specify this parameter, the stream uses USER_PARTITION_KEY.
None
{'StreamDescriptionSummary': {'RecordDistributionStrategy': 'AUTO | '
'USER_PARTITION_KEY'}}
Provides a summarized description of the specified Kinesis data stream without the shard list.
The information returned includes the stream name, Amazon Resource Name (ARN), status, record retention period, approximate creation time, monitoring, encryption details, and open shard count.
DescribeStreamSummary has a limit of 20 transactions per second per account.
See also: AWS API Documentation
Request Syntax
client.describe_stream_summary(
StreamName='string',
StreamARN='string',
StreamId='string'
)
string
The name of the stream to describe.
string
The ARN of the stream.
string
Not Implemented. Reserved for future use.
dict
Response Syntax
{
'StreamDescriptionSummary': {
'StreamName': 'string',
'StreamARN': 'string',
'StreamId': 'string',
'StreamStatus': 'CREATING'|'DELETING'|'ACTIVE'|'UPDATING',
'StreamModeDetails': {
'StreamMode': 'PROVISIONED'|'ON_DEMAND'
},
'RetentionPeriodHours': 123,
'StreamCreationTimestamp': datetime(2015, 1, 1),
'EnhancedMonitoring': [
{
'ShardLevelMetrics': [
'IncomingBytes'|'IncomingRecords'|'OutgoingBytes'|'OutgoingRecords'|'WriteProvisionedThroughputExceeded'|'ReadProvisionedThroughputExceeded'|'IteratorAgeMilliseconds'|'ALL',
]
},
],
'EncryptionType': 'NONE'|'KMS',
'KeyId': 'string',
'OpenShardCount': 123,
'ConsumerCount': 123,
'WarmThroughput': {
'TargetMiBps': 123,
'CurrentMiBps': 123
},
'MaxRecordSizeInKiB': 123,
'ChannelCount': 123,
'RecordDistributionStrategy': 'AUTO'|'USER_PARTITION_KEY'
}
}
Response Structure
(dict) --
StreamDescriptionSummary (dict) --
A StreamDescriptionSummary containing information about the stream.
StreamName (string) --
The name of the stream being described.
StreamARN (string) --
The Amazon Resource Name (ARN) for the stream being described.
StreamId (string) --
Not Implemented. Reserved for future use.
StreamStatus (string) --
The current status of the stream being described. The stream status is one of the following states:
CREATING - The stream is being created. Kinesis Data Streams immediately returns and sets StreamStatus to CREATING.
DELETING - The stream is being deleted. The specified stream is in the DELETING state until Kinesis Data Streams completes the deletion.
ACTIVE - The stream exists and is ready for read and write operations or deletion. You should perform read and write operations only on an ACTIVE stream.
UPDATING - Shards in the stream are being merged or split. Read and write operations continue to work while the stream is in the UPDATING state.
StreamModeDetails (dict) --
Specifies the capacity mode to which you want to set your data stream. Currently, in Kinesis Data Streams, you can choose between an on-demand ycapacity mode and a provisioned capacity mode for your data streams.
StreamMode (string) --
Specifies the capacity mode to which you want to set your data stream. Currently, in Kinesis Data Streams, you can choose between an on-demand capacity mode and a provisioned capacity mode for your data streams.
RetentionPeriodHours (integer) --
The current retention period, in hours.
StreamCreationTimestamp (datetime) --
The approximate time that the stream was created.
EnhancedMonitoring (list) --
Represents the current enhanced monitoring settings of the stream.
(dict) --
Represents enhanced metrics types.
ShardLevelMetrics (list) --
List of shard-level metrics.
The following are the valid shard-level metrics. The value " ALL" enhances every metric.
IncomingBytes
IncomingRecords
OutgoingBytes
OutgoingRecords
WriteProvisionedThroughputExceeded
ReadProvisionedThroughputExceeded
IteratorAgeMilliseconds
ALL
For more information, see Monitoring the Amazon Kinesis Data Streams Service with Amazon CloudWatch in the Amazon Kinesis Data Streams Developer Guide.
(string) --
EncryptionType (string) --
The encryption type used. This value is one of the following:
KMS
NONE
KeyId (string) --
The GUID for the customer-managed Amazon Web Services KMS key to use for encryption. This value can be a globally unique identifier, a fully specified ARN to either an alias or a key, or an alias name prefixed by "alias/".You can also use a master key owned by Kinesis Data Streams by specifying the alias aws/kinesis.
Key ARN example: arn:aws:kms:us-east-1:123456789012:key/12345678-1234-1234-1234-123456789012
Alias ARN example: arn:aws:kms:us-east-1:123456789012:alias/MyAliasName
Globally unique key ID example: 12345678-1234-1234-1234-123456789012
Alias name example: alias/MyAliasName
Master key owned by Kinesis Data Streams: alias/aws/kinesis
OpenShardCount (integer) --
The number of open shards in the stream.
ConsumerCount (integer) --
The number of enhanced fan-out consumers registered with the stream.
WarmThroughput (dict) --
The warm throughput in MB/s for the stream. This represents the throughput capacity that will be immediately available for write operations.
TargetMiBps (integer) --
The target warm throughput value on the stream. This indicates that the stream is currently scaling towards this target value.
CurrentMiBps (integer) --
The current warm throughput value on the stream. This is the write throughput in MiBps that the stream is currently scaled to handle.
MaxRecordSizeInKiB (integer) --
The maximum record size of a single record in kibibyte (KiB) that you can write to, and read from a stream.
ChannelCount (integer) --
The number of channels associated with the stream.
RecordDistributionStrategy (string) --
The record distribution strategy that the stream currently uses. A value of AUTO indicates that Amazon Kinesis Data Streams distributes records across shards using service-managed algorithms. A value of USER_PARTITION_KEY indicates that shard placement is determined by the partition key that producers supply. This field is only present for streams that use the on-demand capacity mode.