Processing CDC Debezium streams

Debezium is a change data capture service for streaming DB changes to other systems for processing. With Yandex Data Streams, you can capture these changes and send them to Cloud Functions.

Below is the architecture of the solution:

Setup

To get a data stream:

1. Create a Yandex Data Streams stream.
2. Set Yandex Data Streams connection parameters.
3. Set up and start the Debezium Server.
4. Set up a trigger in Cloud Functions for processing data.

Creating a stream

Create a Yandex Data Streams stream named debezium. For detailed information about creating streams, see the Yandex Data Streams documentation.

Setting Yandex Data Streams connection parameters

1. Log in to the management console. If not signed up yet, navigate to the management console and follow the on-screen instructions.
2. On the Yandex Cloud Billing page, make sure you have a billing account linked and its status is ACTIVE or TRIAL_ACTIVE. If you do not have a billing account yet, create one.
3. If you do not have a folder yet, create one.
4. Create a service account and assign it the editor role for your folder.
5. Create a static access key.
6. Configure the AWS CLI:

   1. Install the AWS CLI and run this command:

      aws configure
      

   2. Enter the following one by one:

      • AWS Access Key ID [None]:: Service account key ID.
      • AWS Secret Access Key [None]:: Service account secret key.
      • Default region name [None]:: ru-central1 availability zone.

Setting up Debezium Server

This use case shows how Debezium works with PostgreSQL. Let's assume that Debezium is installed on the server where PostgreSQL is running.

1. Install the Debezium Server by following this guide.

2. Go to the conf folder and create the file application.properties with the following content:

   debezium.sink.type=kinesis
   debezium.sink.kinesis.region=ru-central1
   debezium.sink.kinesis.endpoint=<endpoint>
   debezium.source.connector.class=io.debezium.connector.postgresql.PostgresConnector
   debezium.source.offset.storage.file.filename=data/offsets.dat
   debezium.source.offset.flush.interval.ms=0
   debezium.source.database.hostname=localhost
   debezium.source.database.port=5432
   debezium.source.database.user=<username>
   debezium.source.database.password=<user_password>
   debezium.source.database.dbname=<DB_name>
   debezium.source.database.server.name=debezium
   debezium.source.plugin.name=pgoutput
   
   debezium.source.transforms=Reroute
   debezium.source.transforms.Reroute.type=io.debezium.transforms.ByLogicalTableRouter
   debezium.source.transforms.Reroute.topic.regex=(.*)
   debezium.source.transforms.Reroute.topic.replacement=<data_stream>
   

   Where:

   • <endpoint>: Endpoint for a Data Streams data stream, e.g., https://yds.serverless.yandexcloud.net/ru-central1/b1g89ae43m6he********/etn01eg4rn1********. You can find the endpoint on the stream page (see Viewing a list of streams).
   • <data_stream>: Data Streams data stream name.
   • <DB_name>: PostgreSQL database name.
   • <username>: Username for connecting to the PostgreSQL database.
   • <user_password>: User password for connecting to the PostgreSQL database.

3. Run Debezium using this command:

   JAVA_OPTS=-Daws.cborEnabled=false ./run.sh
   

4. Make some changes to the PostgreSQL database, such as insert data into a table.

5. If the settings are correct, the Debezium console will show messages like:

   2022-02-11 07:31:12,850 INFO  [io.deb.con.com.BaseSourceTask](pool-7-thread-1) 1 records sent during previous 00:19:59.999, last recorded offset: {transaction_id=null, lsn_proc=23576408, lsn_commit=23576120, lsn=23576408, txId=580, ts_usec=1644564672582666}
   

Setting up a trigger in Cloud Functions

Create a trigger in Cloud Functions to the Yandex Data Streams debezium stream you created earlier.

For detailed information about creating triggers, see the Cloud Functions documentation.

Cloud Functions will send notifications about any DB changes to the trigger. In the trigger code, you can handle these changes by implementing any required software processing.