CREATE SOURCE: Kafka

CREATE SOURCE connects Materialize to an external system you want to read data from, and provides details about how to decode and interpret that data.

To connect to a Kafka broker (and optionally a schema registry), you first need to create a connection that specifies access and authentication parameters. Once created, a connection is reusable across multiple CREATE SOURCE and CREATE SINK statements.

Syntax

format_spec

key_strat

val_strat

strat

with_options

CONNECTION options

WITH options

No options yet.

Supported formats

Key-value encoding

By default, the message key is decoded using the same format as the message value. However, you can set the key and value encodings explicitly using the KEY FORMAT ... VALUE FORMAT syntax.

Features

Handling upserts

To create a source that uses the standard key-value convention to support inserts, updates, and deletes within Materialize, you can use ENVELOPE UPSERT:

CREATE SOURCE kafka_upsert
  FROM KAFKA CONNECTION kafka_connection (TOPIC 'events')
  FORMAT AVRO USING CONFLUENT SCHEMA REGISTRY CONNECTION csr_connection
  ENVELOPE UPSERT;

Note that:

• Using this envelope is required to consume log compacted topics.
• This envelope can lead to high memory and disk utilization in the cluster maintaining the source. We recommend using a standard-sized cluster, rather than a legacy-sized cluster, to automatically spill the workload to disk. See spilling to disk for details.

Null keys

If a message with a NULL key is detected, Materialize sets the source into an error state. To recover an errored source, you must produce a record with a NULL value and a NULL key to the topic, to force a retraction.

As an example, you can use kcat to produce an empty message:

echo ":" | kcat -b $BROKER -t $TOPIC -Z -K: \
  -X security.protocol=SASL_SSL \
  -X sasl.mechanisms=SCRAM-SHA-256 \
  -X sasl.username=$KAFKA_USERNAME \
  -X sasl.password=$KAFKA_PASSWORD

Using Debezium

Materialize provides a dedicated envelope (ENVELOPE DEBEZIUM) to decode Kafka messages produced by Debezium. To create a source that interprets Debezium messages:

CREATE SOURCE kafka_repl
  FROM KAFKA CONNECTION kafka_connection (TOPIC 'pg_repl.public.table1')
  FORMAT AVRO USING CONFLUENT SCHEMA REGISTRY CONNECTION csr_connection
  ENVELOPE DEBEZIUM;

Any materialized view defined on top of this source will be incrementally updated as new change events stream in through Kafka, as a result of INSERT, UPDATE and DELETE operations in the original database.

For more details and a step-by-step guide on using Kafka+Debezium for Change Data Capture (CDC), check Using Debezium.

Note that:

• This envelope can lead to high memory utilization in the cluster maintaining the source. To reduce memory utilization, consider enabling spill to disk.

Spilling to disk

Kafka sources that use ENVELOPE UPSERT or ENVELOPE DEBEZIUM require storing the current value for each key in the source to produce retractions when keys are updated. When using standard cluster sizes, Materialize will automatically offload this state to disk, seamlessly handling key spaces that are larger than memory.

Spilling to disk is not available with legacy cluster sizes.

Exposing source metadata

In addition to the message value, Materialize can expose the message key, headers and other source metadata fields to SQL.

Key

The message key is exposed via the INCLUDE KEY option. Composite keys are also supported (#7645).

CREATE SOURCE kafka_metadata
  FROM KAFKA CONNECTION kafka_connection (TOPIC 'data')
  KEY FORMAT TEXT
  VALUE FORMAT TEXT
  INCLUDE KEY AS renamed_id;

Note that:

• This option requires specifying the key and value encodings explicitly using the KEY FORMAT ... VALUE FORMAT syntax.

• The UPSERT envelope always includes keys.

• The DEBEZIUM envelope is incompatible with this option.

Headers

Message headers can be retained in Materialize and exposed as part of the source data.

Note that:

• The DEBEZIUM envelope is incompatible with this option.

All headers

All of a message’s headers can be exposed using INCLUDE HEADERS, followed by an AS <header_col>.

This introduces column with the name specified or headers if none was specified. The column has the type record(key: text, value: bytea?) list, i.e. a list of records containing key-value pairs, where the keys are text and the values are nullable byteas.

CREATE SOURCE kafka_metadata
  FROM KAFKA CONNECTION kafka_connection (TOPIC 'data')
  FORMAT AVRO USING CONFLUENT SCHEMA REGISTRY CONNECTION csr_connection
  INCLUDE HEADERS
  ENVELOPE NONE;

To simplify turning the headers column into a map (so individual headers can be searched), you can use the map_build function, e.g.

SELECT
    id,
    seller,
    item,
    convert_from(map_build(headers)->'client_id', 'utf-8') AS client_id,
    map_build(headers)->'encryption_key' AS encryption_key,
FROM kafka_metadata;

 id | seller |        item        | client_id |    encryption_key
----+--------+--------------------+-----------+----------------------
  2 |   1592 | Custom Art         |        23 | \x796f75207769736821
  3 |   1411 | City Bar Crawl     |        42 | \x796f75207769736821

Individual headers

Individual message headers can be exposed via the INCLUDE HEADER key AS name option.

The bytea value of the header is automatically parsed into an UTF-8 string. To expose the raw bytea instead, the BYTES option can be used.

CREATE SOURCE kafka_metadata
  FROM KAFKA CONNECTION kafka_connection (TOPIC 'data')
  FORMAT AVRO USING CONFLUENT SCHEMA REGISTRY CONNECTION csr_connection
  INCLUDE HEADER 'c_id' AS client_id, HEADER 'key' AS encryption_key BYTES,
  ENVELOPE NONE;

Headers can be queried as any other column in the source:

SELECT
    id,
    seller,
    item,
    client_id::numeric,
    encryption_key
FROM kafka_metadata;

 id | seller |        item        | client_id |    encryption_key
----+--------+--------------------+-----------+----------------------
  2 |   1592 | Custom Art         |        23 | \x796f75207769736821
  3 |   1411 | City Bar Crawl     |        42 | \x796f75207769736821

Note that:

• Messages that do not contain all header keys as specified in the source DDL will cause an error that prevents further querying the source.
• Header values containing badly formed UTF-8 strings will cause an error in the source that prevents querying it, unless the BYTES option is specified.

Partition, offset, timestamp

These metadata fields are exposed via the INCLUDE PARTITION, INCLUDE OFFSET and INCLUDE TIMESTAMP options.

CREATE SOURCE kafka_metadata
  FROM KAFKA CONNECTION kafka_connection (TOPIC 'data')
  FORMAT AVRO USING CONFLUENT SCHEMA REGISTRY CONNECTION csr_connection
  INCLUDE PARTITION, OFFSET, TIMESTAMP AS ts
  ENVELOPE NONE;

SELECT "offset" FROM kafka_metadata WHERE ts > '2021-01-01';

offset
------
15
14
13

Setting start offsets

To start consuming a Kafka stream from a specific offset, you can use the START OFFSET option.

CREATE SOURCE kafka_offset
  FROM KAFKA CONNECTION kafka_connection (
    TOPIC 'data',
    -- Start reading from the earliest offset in the first partition,
    -- the second partition at 10, and the third partition at 100.
    START OFFSET (0, 10, 100)
  )
  FORMAT AVRO USING CONFLUENT SCHEMA REGISTRY CONNECTION csr_connection;

Note that:

• If fewer offsets than partitions are provided, the remaining partitions will start at offset 0. This is true if you provide START OFFSET (1) or START OFFSET (1, ...).
• Providing more offsets than partitions is not supported.

Time-based offsets

It’s also possible to set a start offset based on Kafka timestamps, using the START TIMESTAMP option. This approach sets the start offset for each available partition based on the Kafka timestamp and the source behaves as if START OFFSET was provided directly.

It’s important to note that START TIMESTAMP is a property of the source: it will be calculated once at the time the CREATE SOURCE statement is issued. This means that the computed start offsets will be the same for all views depending on the source and stable across restarts.

If you need to limit the amount of data maintained as state after source creation, consider using temporal filters instead.

CONNECTION options

KEY STRATEGY and VALUE STRATEGY

It is possible to define how an Avro reader schema will be chosen for Avro sources by using the KEY STRATEGY and VALUE STRATEGY keywords, as shown in the syntax diagram.

A strategy of LATEST (the default) will choose the latest writer schema from the schema registry to use as a reader schema. ID or INLINE will allow specifying a schema from the registry by ID or inline in the CREATE SOURCE statement, respectively.

Monitoring source progress

By default, Kafka sources expose progress metadata as a subsource that you can use to monitor source ingestion progress. The name of the progress subsource can be specified when creating a source using the EXPOSE PROGRESS AS clause; otherwise, it will be named <src_name>_progress.

The following metadata is available for each source as a progress subsource:

And can be queried using:

SELECT
  partition, "offset"
FROM
  (
    SELECT
      -- Take the upper of the range, which is null for non-partition rows
      -- Cast partition to u64, which is more ergonomic
      upper(partition)::uint8 AS partition, "offset"
    FROM
      <src_name>_progress
  )
WHERE
  -- Remove all non-partition rows
  partition IS NOT NULL;

As long as any offset continues increasing, Materialize is consuming data from the upstream Kafka broker. For more details on monitoring source ingestion progress and debugging related issues, see Troubleshooting.

Monitoring consumer lag

To support Kafka tools that monitor consumer lag, Kafka sources commit offsets once the messages up through that offset have been durably recorded in Materialize’s storage layer.

However, rather than relying on committed offsets, Materialize suggests using our native progress monitoring, which contains more up-to-date information.

Committed offsets are associated with a consumer group specific to the source. The ID of the consumer group consists of the prefix configured with the GROUP ID PREFIX option followed by a Materialize-generated suffix.

You should not make assumptions about the number of consumer groups that Materialize will use to consume from a given source. The only guarantee is that the ID of each consumer group will begin with the configured prefix.

The consumer group ID prefix for each Kafka source in the system is available in the group_id_prefix column of the mz_kafka_sources table. To look up the group_id_prefix for a source by name, use:

SELECT group_id_prefix
FROM mz_internal.mz_kafka_sources ks
JOIN mz_sources s ON s.id = ks.id
WHERE s.name = '<src_name>'

Required permissions

The access control lists (ACLs) on the Kafka cluster must allow Materialize to perform the following operations on the following resources:

Examples

Creating a connection

A connection describes how to connect and authenticate to an external system you want Materialize to read data from.

Once created, a connection is reusable across multiple CREATE SOURCE statements. For more details on creating connections, check the CREATE CONNECTION documentation page.

Broker

CREATE SECRET kafka_ssl_key AS '<BROKER_SSL_KEY>';
CREATE SECRET kafka_ssl_crt AS '<BROKER_SSL_CRT>';

CREATE CONNECTION kafka_connection TO KAFKA (
    BROKER 'rp-f00000bar.data.vectorized.cloud:30365',
    SSL KEY = SECRET kafka_ssl_key,
    SSL CERTIFICATE = SECRET kafka_ssl_crt
);

CREATE SECRET kafka_password AS '<BROKER_PASSWORD>';

CREATE CONNECTION kafka_connection TO KAFKA (
    BROKER 'unique-jellyfish-0000-kafka.upstash.io:9092',
    SASL MECHANISMS = 'SCRAM-SHA-256',
    SASL USERNAME = 'foo',
    SASL PASSWORD = SECRET kafka_password
);

If your Kafka broker is not exposed to the public internet, you can tunnel the connection through an AWS PrivateLink service or an SSH bastion host:

CREATE CONNECTION privatelink_svc TO AWS PRIVATELINK (
    SERVICE NAME 'com.amazonaws.vpce.us-east-1.vpce-svc-0e123abc123198abc',
    AVAILABILITY ZONES ('use1-az1', 'use1-az4')
);

CREATE CONNECTION kafka_connection TO KAFKA (
    BROKERS (
        'broker1:9092' USING AWS PRIVATELINK privatelink_svc,
        'broker2:9092' USING AWS PRIVATELINK privatelink_svc (PORT 9093)
    )
);

CREATE CONNECTION ssh_connection TO SSH TUNNEL (
    HOST '<SSH_BASTION_HOST>',
    USER '<SSH_BASTION_USER>',
    PORT <SSH_BASTION_PORT>
);

CREATE CONNECTION kafka_connection TO KAFKA (
BROKERS (
    'broker1:9092' USING SSH TUNNEL ssh_connection,
    'broker2:9092' USING SSH TUNNEL ssh_connection
    )
);

Confluent Schema Registry

CREATE SECRET csr_ssl_crt AS '<CSR_SSL_CRT>';
CREATE SECRET csr_ssl_key AS '<CSR_SSL_KEY>';
CREATE SECRET csr_password AS '<CSR_PASSWORD>';

CREATE CONNECTION csr_connection TO CONFLUENT SCHEMA REGISTRY (
    URL 'https://rp-f00000bar.data.vectorized.cloud:30993',
    SSL KEY = SECRET csr_ssl_key,
    SSL CERTIFICATE = SECRET csr_ssl_crt,
    USERNAME = 'foo',
    PASSWORD = SECRET csr_password
);

CREATE SECRET IF NOT EXISTS csr_username AS '<CSR_USERNAME>';
CREATE SECRET IF NOT EXISTS csr_password AS '<CSR_PASSWORD>';

CREATE CONNECTION csr_connection TO CONFLUENT SCHEMA REGISTRY (
  URL '<CONFLUENT_REGISTRY_URL>',
  USERNAME = SECRET csr_username,
  PASSWORD = SECRET csr_password
);

If your Confluent Schema Registry server is not exposed to the public internet, you can tunnel the connection through an AWS PrivateLink service or an SSH bastion host:

CREATE CONNECTION privatelink_svc TO AWS PRIVATELINK (
    SERVICE NAME 'com.amazonaws.vpce.us-east-1.vpce-svc-0e123abc123198abc',
    AVAILABILITY ZONES ('use1-az1', 'use1-az4')
);

CREATE CONNECTION csr_connection TO CONFLUENT SCHEMA REGISTRY (
    URL 'http://my-confluent-schema-registry:8081',
    AWS PRIVATELINK privatelink_svc
);

CREATE CONNECTION ssh_connection TO SSH TUNNEL (
    HOST '<SSH_BASTION_HOST>',
    USER '<SSH_BASTION_USER>',
    PORT <SSH_BASTION_PORT>
);

CREATE CONNECTION csr_connection TO CONFLUENT SCHEMA REGISTRY (
    URL 'http://my-confluent-schema-registry:8081',
    SSH TUNNEL ssh_connection
);

Creating a source

CREATE SOURCE avro_source
  FROM KAFKA CONNECTION kafka_connection (TOPIC 'test_topic')
  FORMAT AVRO USING CONFLUENT SCHEMA REGISTRY CONNECTION csr_connection;

CREATE SOURCE json_source
  FROM KAFKA CONNECTION kafka_connection (TOPIC 'test_topic')
  FORMAT JSON;

CREATE VIEW typed_kafka_source AS
  SELECT
    (data->>'field1')::boolean AS field_1,
    (data->>'field2')::int AS field_2,
    (data->>'field3')::float AS field_3
  FROM json_source;

CREATE SOURCE proto_source
  FROM KAFKA CONNECTION kafka_connection (TOPIC 'test_topic')
  FORMAT PROTOBUF USING CONFLUENT SCHEMA REGISTRY CONNECTION csr_connection;

CREATE SOURCE text_source
  FROM KAFKA CONNECTION kafka_connection (TOPIC 'test_topic')
  FORMAT TEXT
  ENVELOPE UPSERT;

CREATE SOURCE csv_source (col_foo, col_bar, col_baz)
  FROM KAFKA CONNECTION kafka_connection (TOPIC 'test_topic')
  FORMAT CSV WITH 3 COLUMNS;

Related pages

• CREATE SECRET
• CREATE CONNECTION
• CREATE SOURCE
• SHOW SOURCES
• DROP SOURCE
• Using Debezium