Message Filtering

Table of Contents

Set up
Example: Basic configuration
Variables for use in filter expressions
Language specifics
Configuration options

By default, Debezium delivers every data change event that it receives to the Kafka broker. However, in many cases, you might be interested in only a subset of the events emitted by the producer. To enable you to process only the records that are relevant to you, Debezium provides the filter single message transform (SMT).

The filter SMT is under active development. The structure of the emitted message or other details might change as development progresses.

While it is possible to use Java to create a custom SMT to encode filtering logic, using a custom-coded SMT has its drawbacks. For example:

It is necessary to compile the transformation up front and deploy it to Kafka Connect.
Every change needs code recompilation and redeployment, leading to inflexible operations.

The filter SMT supports scripting languages that integrate with JSR 223 (Scripting for the Java™ Platform).

Debezium does not come with any implementations of the JSR 223 API. To use an expression language with Debezium, you must download the JSR 223 script engine implementation for the language, and add to your Debezium connector plug-in directories, along any other JAR files used by the language implementation. For example, for Groovy 3, you can download its JSR 223 implementation from https://groovy-lang.org/. The JSR223 implementation for GraalVM JavaScript is available at https://github.com/graalvm/graaljs.

Set up

For security reasons, the filter SMT is not included with the Debezium connector archives. Instead, it is provided in a separate artifact, debezium-scripting-1.4.2.Final.tar.gz. To use the filter SMT with a Debezium connector plug-in, you must explicitly add the SMT artifact to your Kafka Connect environment.

After the filter SMT is present in a Kafka Connect instance, any user who is allowed to add a connector to the instance can run scripting expressions. To ensure that scripting expressions can be run only by authorized users, be sure to secure the Kafka Connect instance and its configuration interface before you add the filter SMT.

With Zookeeper, Kafka, Kafka Connect, and one or more Debezium connectors installed, the remaining tasks to install the filter SMT are:

Download the scripting SMT archive
Extract the contents of the archive into the Debezium plug-in directories of your Kafka Connect environment.
Obtain a JSR-223 script engine implementation and add its contents to the Debezium plug-in directories of your Kafka Connect environment.
Restart your Kafka Connect process to pick up the new JAR files.

Example: Basic configuration

You configure the filter transformation in the Debezium connector’s Kafka Connect configuration. In the configuration, you specify the events that you are interested in by defining filter conditions that are based on business rules. As the filter SMT processes the event stream, it evaluates each event against the configured filter conditions. Only events that meet the criteria of the filter conditions are passed to the broker.

To configure a Debezium connector to filter change event records, configure the Filter SMT in the Kafka Connect configuration for the Debezium connector. Configuration of the filter SMT requires you to specify a regular expression that defines the filtering criteria.

For example, you might add the following configuration in your connector configuration.

...
transforms=filter
transforms.filter.type=io.debezium.transforms.Filter
transforms.filter.language=jsr223.groovy
transforms.filter.condition=value.op == 'u' && value.before.id == 2
...

The preceding example specifies the use of the Groovy expression language. The regular expression value.op == 'u' && value.before.id == 2 removes all messages, except those that represent update (u) records with id values that are equal to 2.

Variables for use in filter expressions

Debezium binds certain variables into the evaluation context for the filter SMT. When you create expressions to specify filter conditions, you can use the variables that Debezium binds into the evaluation context. By binding variables, Debezium enables the SMT to look up and interpret their values as it evaluates the conditions in an expression.

The following table lists the variables that Debezium binds into the evaluation context for the filter SMT:

Table 1. Filter expression variables
Name	Description	Type
`key`	A key of the message.	`org.apache.kafka.connect.data.Struct`
`value`	A value of the message.	`org.apache.kafka.connect.data.Struct`
`keySchema`	Schema of the message key.	`org.apache.kafka.connect.data.Schema`
`valueSchema`	Schema of the message value.	`org.apache.kafka.connect.data.Schema`
`topic`	Name of the target topic.	String
`headers`	A Java map of message headers. The key field is the header name. The `headers` variable exposes the following properties: `value` (of type `Object`) `schema` (of type `org.apache.kafka.connect.data.Schema`)	`java.util.Map<String, io.debezium.transforms.scripting.RecordHeader>`

An expression can invoke arbitrary methods on its variables. Expressions should resolve to a Boolean value that determines how the SMT dispositions the message. When the filter condition in an expression evaluates to true, the message is retained. When the filter condition evaluates to false, the message is removed.

Expressions should not result in any side-effects. That is, they should not modify any variables that they pass.

Language specifics

The way that you express filtering conditions depends on the scripting language that you use.

For example, as shown in the basic configuration example, when you use Groovy as the expression language, the following expression removes all messages, except for update records that have id values set to 2:

value.op == 'u' && value.before.id == 2

Other languages use different methods to express the same condition.

The Debezium MongoDB connector emits the after and patch fields as serialized JSON documents rather than as structures. To use the filter SMT with the MongoDB connector, you must first unwind the fields by applying the ExtractNewDocumentState SMT.

You could also take the approach of using a JSON parser within the expression. For example, if you use Groovy as the expression language, add the groovy-json artifact to the classpath, and then add an expression such as (new groovy.json.JsonSlurper()).parseText(value.after).last_name == 'Kretchmar'.

Javascript

If you use JavaScript as the expression language, you can call the Struct#get() method to specify the filtering condition, as in the following example:

value.get('op') == 'u' && value.get('before').get('id') == 2

Javascript with Graal.js

If you use JavaScript with Graal.js to define filtering conditions, you use an approach that is similar to the one that you use with Groovy. For example:

value.op == 'u' && value.before.id == 2

Configuration options

The following table lists the configuration options that you can use with the filter SMT.

Table 2. filter SMT configuration options
Property	Default	Description
`topic.regex`		An optional regular expression that evaluates the name of the destination topic for an event to determine whether to apply filtering logic. If the name of the destination topic matches the value in `topic.regex`, the transformation applies the filter logic before it passes the event to the topic. If the name of the topic does not match the value in `topic.regex`, the SMT passes the event to the topic unmodified.
`language`		The language in which the expression is written. Must begin with `jsr223.`, for example, `jsr223.groovy`, or `jsr223.graal.js`. Debezium supports bootstrapping through the JSR 223 API ("Scripting for the Java ™ Platform") only.
`condition`		The expression to be evaluated for every message. Must evaluate to a Boolean value where a result of `true` keeps the message, and a result of `false` removes it.
`null.handling.mode`	`keep`	Specifies how the transformation handles `null` (tombstone) messages. You can specify one of the following options: `keep` (Default) Pass the messages through. `drop` Remove the messages completely. `evaluate` Apply the filter condition to the messages.