kafka connect 大概了解

news2025/12/18 5:35:14

kafka connect

Introduction

Kafka Connect is the component of Kafka that provides data integration between databases, key-value stores, search indexes, file systems, and Kafka brokers.

kafka connect 是一个框架，用来帮助集成其他系统的数据到kafka，或者将kafka 的数据吐到其他系统，例如数据库，Elasticsearch 之类的外部系统。

在kafka connect 里，有 kafka woker, kafka connector，kafka connect plugin的概念。

一个kafka worker 就是一个instance，类似一个pod 这种独立的实例。

kafka connector 就是用来搬运数据的连接器。有 source connector 和 sink connector.

Source 就是数据来源的connector，sink 就是吐出数据的 connector.

例如

Connector Type	Name	Purpose
Source	`JdbcSourceConnector`	Pull data from relational DBs
Sink	`ElasticsearchSinkConnector`	Push Kafka data to Elasticsearch
Source	`FileStreamSourceConnector`	Read lines from file into Kafka

connector 简略配置

{
  "name": "my-jdbc-source",
  "connector.class": "io.confluent.connect.jdbc.JdbcSourceConnector",
  "tasks.max": "1",
  "connection.url": "jdbc:mysql://localhost:3306/mydb",
  "topic.prefix": "mysql-"
}

A Kafka Connect Plugin is the packaged code (JAR files) that implements one or more connectors.

It’s usually installed by placing the plugin into Kafka Connect’s plugin.path directory.

A plugin might include:

The connector logic
Converters (e.g., JSON, Avro)
Transformations (optional logic to modify data)

Think of a connector as a configuration, and a plugin as the actual implementation that makes it work.

实现结构

从部署的角度来看，kafka connect 是一个独立的service cluster。
下面的docker-compose.yml 配置可以看出 cp-kafka-connect 这个image 就可可以load 在 CONNECT_PLUGIN_PATH 配置目录下的 connect plugin 来实现不同的 connector 功能。

而 confluent-hub install --no-prompt debezium/debezium-connector-mysql:1.7.1 这个命令就是部署mysql 相关的source connector 到 CONNECT_PLUGIN_PATH 配置的目录usr/share/confluent-hub-components 下面。

kafka-connect:
  image: confluentinc/cp-kafka-connect:7.1.0-1-ubi8
  environment:
    CONNECT_PLUGIN_PATH: /usr/share/java,/usr/share/confluent-hub-components

  command:
    - bash
    - -c
    - |
      confluent-hub install --no-prompt debezium/debezium-connector-mysql:1.7.1
      /etc/confluent/docker/run