Sinks¶

Tinybird Sinks allow you to export data from your Tinybird Workspace to external systems on a scheduled or on-demand basis. Sinks are built on top of Pipes and provide a fully managed way to push data to various destinations.

Available Sinks¶

Tinybird supports the following Sink destinations:

• Kafka Sink
• S3 Sink
• GCS Sink

Create a Sink¶

Sinks are defined with a connection and a pipe that exports query results to that connection. Use TYPE SINK in .pipe files, defineSinkPipe in the TypeScript SDK, or define_sink_pipe in the Python SDK.

The following example creates an S3 Sink that exports processed events to an S3 bucket on a schedule.

TYPE s3
S3_REGION "us-east-1"
S3_ARN {{ tb_secret("AWS_ROLE_ARN") }}

SCHEMA >
    `timestamp` DateTime,
    `session_id` String,
    `status` String

ENGINE "MergeTree"
ENGINE_SORTING_KEY "timestamp"

NODE export
SQL >
    SELECT timestamp, session_id
    FROM events
    WHERE status = 'processed'

TYPE SINK
EXPORT_CONNECTION_NAME "landing_s3"
EXPORT_BUCKET_URI "s3://my-bucket/exports/"
EXPORT_FILE_TEMPLATE "events_{date}"
EXPORT_FORMAT "csv"
EXPORT_SCHEDULE "0 * * * *"

See all syntax options in the Sink pipes reference, TypeScript SDK reference, and Python SDK reference.

Key features¶

• Fully Managed: Sinks require no additional tooling or infrastructure management.
• Scheduled or On-Demand: Run exports on a defined schedule using cron expressions or trigger them manually when needed.
• Query Parameters: Support for parameterized queries allows flexible data filtering and transformation.
• Observability: Monitor Sink operations and data transfer through Service Data Sources.

Common use cases¶

Sinks enable various data integration scenarios:

• Regular data exports to clients or partner systems.
• Feeding data lakes and data warehouses.
• Real-time data synchronization with external systems.
• Event-driven architectures and data pipelines.