SKILL.md
Amazon Kinesis Data Streams
Overview
Domain expertise for Amazon Kinesis Data Streams. This skill covers streaming tables and Amazon S3 delivery, which provide serverless, fully managed delivery from a KDS stream directly to Apache Iceberg tables on S3 Tables or to general-purpose S3 buckets — no consumers, no Firehose, no Flink required for append-only delivery.
Delivery is configured as a channel on a stream via CreateChannel. Channels consume no shard capacity and no enhanced fan-out slots, so they do not compete with existing consumers.
The AWS MCP server is recommended for executing AWS operations. When the MCP server is not available, use the AWS CLI or shell commands instead.
Which Workflow Do You Need?
| Customer Intent | Reference |
|---|---|
| Deliver stream data to Apache Iceberg tables on S3 Tables — setup, IAM, Glue schema, type mapping, partitioning, channel lifecycle | references/streaming-tables.md |
| Deliver stream data to a general-purpose S3 bucket as objects — setup, IAM, compression, storage class, output key templates, channel lifecycle | references/general-purpose-delivery-to-s3.md |
| Build a lakehouse / data lake from Kinesis; make streaming data queryable in Athena | references/streaming-tables.md |
| Zero-ops, serverless alternative to Firehose or Kinesis consumers for KDS → S3 delivery | references/general-purpose-delivery-to-s3.md |
| CloudWatch metrics and alarms, delivery logging, CloudTrail, encryption, channel states, quotas, naming rules, what can be updated | references/monitoring-security-and-limits.md |
| Channel is ACTIVE but no data arrives; records landing in the DLQ; rising data freshness | Start with references/monitoring-security-and-limits.md, then the destination-specific troubleshooting table |
| Deliver to both S3 and S3 Tables from one stream | Supported — create one channel per destination |
Prerequisites
Both destinations require an ON_DEMAND stream — provisioned-mode streams are not supported. Each channel also needs an IAM service execution role trusted by kinesis.amazonaws.com, and a destination in the same Region as the stream.
Hard Limits to Check First
These rule out the feature entirely, so confirm them before designing a solution:
- Append-only — no CDC, upserts, or deletes
- No schema evolution — a schema change means deleting and recreating the channel
- No backfill — only records produced after the channel is
ACTIVEare delivered - No transformations — records are delivered as-is
- New table per channel — cannot deliver into an existing Iceberg table
- Data freshness is 300–900 seconds — not sub-minute
- Same Region only — cross-account is supported, cross-Region is not
If any of these block the customer, recommend Managed Service for Apache Flink instead.