Sinks¶

Choose where to deliver your events.

Overview¶

Sinks define where Postgres Stream delivers events. Each sink type has a dedicated Docker image for minimal size.

Available Sinks¶

Sink	Docker Image	Use Case
Kafka	`kafka-latest`	High-throughput event streaming
NATS	`nats-latest`	Lightweight pub/sub messaging
RabbitMQ	`rabbitmq-latest`	Enterprise message broker
Redis Strings	`redis-strings-latest`	Key-value caching
Redis Streams	`redis-streams-latest`	Append-only event log
Webhook	`webhook-latest`	HTTP POST delivery
AWS SQS	`sqs-latest`	Managed queue service
AWS SNS	`sns-latest`	Managed pub/sub service
AWS Kinesis	`kinesis-latest`	Real-time data streaming
GCP Pub/Sub	`gcp-pubsub-latest`	Google Cloud messaging
Elasticsearch	`elasticsearch-latest`	Search indexing
Meilisearch	`meilisearch-latest`	Search indexing

Dynamic Routing¶

Most sinks support dynamic routing via event metadata. This lets you route events to different destinations based on row data:

-- Route to different topics based on table name
insert into pgstream.subscriptions (
  key, stream_id, operation, schema_name, table_name,
  column_names, metadata_extensions
) values (
  'all-events',
  1,
  'INSERT',
  'public',
  'orders',
  array['id', 'user_id'],
  '[{"json_path": "topic", "expression": "''events-'' || tg_table_name"}]'::jsonb
);

See each sink's documentation for supported metadata fields.

Choosing a Sink¶

For high throughput: Kafka, Kinesis, or GCP Pub/Sub

For simple integration: Webhook

For AWS infrastructure: SQS, SNS, or Kinesis

For GCP infrastructure: GCP Pub/Sub

For real-time messaging: NATS or RabbitMQ

For caching: Redis Strings

For event log: Redis Streams

For search indexing: Elasticsearch or Meilisearch