target-kafka 0.1.3

Singer target for Kafka, built with the Meltano Singer SDK.

target-kafka

target-kafka is a Singer target for Apache Kafka, built with the Meltano Singer SDK. It writes tap output to Kafka topics (one topic per stream) with configurable batching, compression, and message keys. Delivery is at-least-once: on failure, records may be retried and duplicates are possible.

Installation

Install from PyPI:

uv tool install target-kafka

Install from GitHub (replace your-org with your GitHub org or username):

uv tool install git+https://github.com/your-org/target-kafka.git@main

Publishing: To publish to PyPI and add this loader to Meltano Hub, see docs/PUBLISHING.md. For version bumps and release steps, see docs/RELEASING.md.

Configuration

Accepted config options

Setting	Required	Default	Description
bootstrap_servers	Yes	—	Kafka bootstrap servers (comma-separated, e.g. localhost:9092,localhost:9093)
topic_prefix	No	""	Prefix to add to all topic names
key_properties	No	[]	Record properties to use as Kafka message key (empty = round-robin partition)
batch_size	No	100	Number of records to batch before sending to Kafka
compression_type	No	snappy	Kafka message compression: none, gzip, snappy, lz4, zstd
include_sdc_properties	No	true	Include Meltano metadata (_sdc_extracted_at, _sdc_received_at, etc.)
common_topic	No	—	When set, all streams write to this topic; each record gets a stream field with the source stream name and validation is disabled. When unset, each stream uses its name as the topic (with optional topic_prefix).

Common topic (all streams to one topic): Set common_topic (e.g. raw-slack-events) to send every stream to that single topic. The target adds a stream property to each record with the source stream name so consumers can filter or route. Schema validation is disabled in this mode because records from different streams may have different shapes. If the tap already sends a stream property, it is overwritten with the target’s stream name.

A full list of supported settings and capabilities is available by running:

target-kafka --about

Configure using environment variables

With --config=ENV, the target reads configuration from environment variables (and from a .env file in the working directory if present). Use names like TARGET_KAFKA_BOOTSTRAP_SERVERS, TARGET_KAFKA_TOPIC_PREFIX, etc. See .env.example for a template.

Usage

Run with Meltano

You can run target-kafka in a pipeline using Meltano:

meltano run tap-smoke-test target-kafka

Run the target directly

target-kafka --version
target-kafka --help

# Example with config file
tap-smoke-test | target-kafka --config config.json

# Example with env config
tap-smoke-test | target-kafka --config ENV

This target works in any Singer environment and does not require Meltano.

Developer resources

Prerequisites

• Python 3.10+
• uv

Setup and tests

uv sync
uv run pytest
uv run target-kafka --help

Testing with Meltano

uv tool install meltano
meltano invoke target-kafka --version
meltano run tap-smoke-test target-kafka

SDK dev guide

See the Singer SDK dev guide for developing taps and targets.