Important

You are viewing documentation for an older version of Confluent Platform. For the latest, click here.

HTTP Sink Connector for Confluent Platform

The Kafka Connect HTTP Sink Connector integrates Apache Kafka® with an API via HTTP or HTTPS.

The connector consumes records from Kafka topic(s) and converts each record value to a String before sending it in the request body to the configured http.api.url, which optionally can reference the record key and/or topic name. The targeted API must support either a POST or PUT request.

The connector batches records up to the set batch.max.size before sending the batched request to the API. Each record is converted to its String representation and then separated with the batch.separator.

The HTTP Sink Connector supports connecting to APIs using SSL along with Basic Authentication, OAuth2, or a Proxy Authentication Server.

Install the HTTP Connector

You can install this connector by using the Confluent Hub client (recommended) or you can manually download the ZIP file.

Install the connector using Confluent Hub

Prerequisite

Confluent Hub Client must be installed. This is installed by default with Confluent Enterprise.

Navigate to your Confluent Platform installation directory and run the following command to install the latest (latest) connector version. The connector must be installed on every machine where Connect will run.

confluent-hub install confluentinc/kafka-connect-http:latest

You can install a specific version by replacing latest with a version number. For example:

confluent-hub install confluentinc/kafka-connect-http:1.0.3

Install Connector Manually

Download and extract the ZIP file for your connector and then follow the manual connector installation instructions.

License

You can use this connector for a 30-day trial period without a license key.

After 30 days, this connector is available under a Confluent enterprise license. Confluent issues enterprise license keys to subscribers, along with providing enterprise-level support for Confluent Platform and your connectors. If you are a subscriber, please contact Confluent Support at support@confluent.io for more information.

See Confluent Platform license for license properties and License topic configuration for information about the license topic.

Quick Start

This quick start uses the HTTP Sink Connector to consume records and send HTTP requests to a demo HTTP service running locally that is running without any authentication.

Additional examples can be found in the Feature Descriptions and Examples section below.

Prerequisites

• Confluent Platform
• Confluent CLI (requires separate installation)

1. Before starting the connector, clone and run the kafka-connect-http-demo app on your machine.

   git clone https://github.com/confluentinc/kafka-connect-http-demo.git
   cd kafka-connect-http-demo
   mvn spring-boot:run -Dspring.profiles.active=simple-auth
   

2. Install the connector through the Confluent Hub Client.

   Tip

   The command syntax for the Confluent CLI development commands changed in 5.3.0. These commands have been moved to confluent local. For example, the syntax for confluent start is now confluent local start. For more information, see confluent local.

   confluent local start
   

3. Produce test data to the http-messages topic in Kafka using the Confluent CLI confluent local produce command.

   seq 10 | confluent local produce http-messages
   

4. Create a http-sink.json file with the following contents:

   {
     "name": "HttpSink",
     "config": {
       "topics": "http-messages",
       "tasks.max": "1",
       "connector.class": "io.confluent.connect.http.HttpSinkConnector",
       "http.api.url": "http://localhost:8080/api/messages",
       "value.converter": "org.apache.kafka.connect.storage.StringConverter",
       "confluent.topic.bootstrap.servers": "localhost:9092",
       "confluent.topic.replication.factor": "1"
     }
   }
   

5. Load the HTTP Sink Connector.

   Caution

   You must include a double dash (--) between the topic name and your flag. For more information, see this post.

   confluent local load httpsink -- -d http-sink.json
   

   Important

   Don’t use the confluent local commands in production environments.

6. Confirm that the connector is in a RUNNING state.

   confluent local status httpsink
   

7. Confirm that the data was sent to the HTTP endpoint.

   curl localhost:8080/api/messages
   

Note

Before running other examples, kill the demo app (CTRL + C) to avoid port conflicts.

Examples

• Authentication
• Header Forwarding
• Key and Topic Substitution
• Regex Replacements
• Tombstone Records
• Retries
• Reporter

Authentication

The HTTP Connector can run with SSL enabled/disabled and also supports various authentication types like Basic Auth, OAuth2, and Proxy Server Auth.

Basic Authentication Example

1. Run the demo app with the basic-auth Spring profile.

   mvn spring-boot:run -Dspring.profiles.active=basic-auth
   

   Note

   If the demo app is already running, you will need to kill that instance (CTRL + C) before running a new instance to avoid port conflicts.

2. Create a http-sink.properties file with the following contents:

   name=HttpSinkBasicAuth
   topics=http-messages
   tasks.max=1
   connector.class=io.confluent.connect.http.HttpSinkConnector
   # key/val converters
   key.converter=org.apache.kafka.connect.storage.StringConverter
   value.converter=org.apache.kafka.connect.storage.StringConverter
   # licensing for local single-node Kafka cluster
   confluent.topic.bootstrap.servers=localhost:9092
   confluent.topic.replication.factor=1
   # http sink connector configs
   http.api.url=http://localhost:8080/api/messages
   auth.type=BASIC
   connection.user=admin
   connection.password=password
   

3. Run and validate the connector as described in the Quick Start.

OAuth2 Authentication Example

Note

The connector’s OAuth2 configuration only allows for use of the Client Credentials grant type.

1. Run the demo app with the oauth2 Spring profile.

   mvn spring-boot:run -Dspring.profiles.active=oauth2
   

2. Create a http-sink.properties file with the following contents:

   name=HttpSinkOAuth2
   topics=http-messages
   tasks.max=1
   connector.class=io.confluent.connect.http.HttpSinkConnector
   # key/val converters
   key.converter=org.apache.kafka.connect.storage.StringConverter
   value.converter=org.apache.kafka.connect.storage.StringConverter
   # licensing for local single-node Kafka cluster
   confluent.topic.bootstrap.servers=localhost:9092
   confluent.topic.replication.factor=1
   # http sink connector configs
   http.api.url=http://localhost:8080/api/messages
   auth.type=OAUTH2
   oauth2.token.url=http://localhost:8080/oauth/token
   oauth2.client.id=kc-client
   oauth2.client.secret=kc-secret
   

3. Run and validate the connector as described in the Quick Start.

SSL with Basic Authentication Example

1. Run the demo app with the ssl-auth Spring profile.

   mvn spring-boot:run -Dspring.profiles.active=ssl-auth
   

2. Create a http-sink.properties file with the following contents:

   name=SSLHttpSink
   topics=string-topic
   tasks.max=1
   connector.class=io.confluent.connect.http.HttpSinkConnector
   # key/val converters
   key.converter=org.apache.kafka.connect.storage.StringConverter
   value.converter=org.apache.kafka.connect.storage.StringConverter
   # licensing for local single-node Kafka cluster
   confluent.topic.bootstrap.servers=localhost:9092
   confluent.topic.replication.factor=1
   # http sink connector configs
   http.api.url=https://localhost:8443/api/messages
   # http sink connector SSL config
   ssl.enabled=true
   https.ssl.truststore.location=/path/to/http-sink-demo/src/main/resources/localhost-keystore.jks
   https.ssl.truststore.type=JKS
   https.ssl.truststore.password=changeit
   https.ssl.keystore.location=/path/to/http-sink-demo/src/main/resources/localhost-keystore.jks
   https.ssl.keystore.type=JKS
   https.ssl.keystore.password=changeit
   https.ssl.key.password=changeit
   https.ssl.protocol=TLSv1.2
   
   auth.type=BASIC
   connection.user=admin
   connection.password=password
   

   Tip

   Don’t forget to update the https.ssl.truststore.location and https.ssl.keystore.location with the path to your http-sink-demo project.

3. Run and validate the connector as described in the Quick Start.

Proxy Authentication Example

Note

The proxy authentication example is dependent on MacOS X 10.6.8 or higher due to the proxy that is utilized.

1. Run the demo app with the simple-auth Spring profile.

   mvn spring-boot:run -Dspring.profiles.active=simple-auth
   

2. Install Squidman Proxy.

3. In Squidman preferences/general, set the http port to 3128.

4. In Squidman preferences/template, add the following:

   auth_param basic program /usr/local/squid/libexec/basic_ncsa_auth /etc/squid/passwords
   auth_param basic realm proxy
   acl authenticated proxy_auth REQUIRED
   http_access allow authenticated
   

5. Create a credentials file for the proxy.

   sudo mkdir /etc/squid
   sudo htpasswd -c /etc/squid/passwords proxyuser
   # set password to proxypassword
   

6. Open the Squidman application and select Start Squid.

7. Create a http-sink.properties file with the following contents:

   name=HttpSinkProxyAuth
   topics=http-messages
   tasks.max=1
   connector.class=io.confluent.connect.http.HttpSinkConnector
   # key/val converters
   key.converter=org.apache.kafka.connect.storage.StringConverter
   value.converter=org.apache.kafka.connect.storage.StringConverter
   # licensing for local single-node Kafka cluster
   confluent.topic.bootstrap.servers=localhost:9092
   confluent.topic.replication.factor=1
   # http sink connector configs
   http.api.url=http://localhost:8080/api/messages
   http.proxy.host=localhost
   http.proxy.port=3128
   http.proxy.user=proxyuser
   http.proxy.password=proxypassword
   

8. Run and validate the connector as described in the Quick Start.

Key and Value Converters

Currently this connector supports StringConverter only.

Header Forwarding

The connector forwards any headers configured via the headers property. Multiple headers can be separated via the | but this is configurable by setting header.separator.

Note

Headers on the incoming Kafka records will not be forwarded.

Header Forwarding Example

1. Run the demo app with the basic-auth Spring profile.

   mvn spring-boot:run -Dspring.profiles.active=basic-auth
   

2. Create a http-sink.properties file with the following contents:

   name=HttpSinkBasicAuth
   topics=http-messages
   tasks.max=1
   connector.class=io.confluent.connect.http.HttpSinkConnector
   # key/val converters
   key.converter=org.apache.kafka.connect.storage.StringConverter
   value.converter=org.apache.kafka.connect.storage.StringConverter
   # licensing for local single-node Kafka cluster
   confluent.topic.bootstrap.servers=localhost:9092
   confluent.topic.replication.factor=1
   # http sink connector configs
   http.api.url=http://localhost:8080/api/messages
   auth.type=BASIC
   connection.user=admin
   connection.password=password
   headers=Forward-Me:header_value|Another-Header:another_value
   

3. Run and validate the connector as described in the Quick Start.

Key and Topic Substitution

The record’s value is the only piece of data forwarded to the API by default. However, the record key and/or topic can be substituted into the http.api.url so that it can be sent to the API.

The example below illustrates how this can be done. Notice the structure of the http.api.url.

Key and Topic Substitution Example

1. Run the demo app with the simple-auth Spring profile.

   mvn spring-boot:run -Dspring.profiles.active=simple-auth
   

2. Create a http-sink.properties file with the following contents:

   name=KeyTopicSubstitution
   topics=key-val-messages
   tasks.max=1
   connector.class=io.confluent.connect.http.HttpSinkConnector
   # key/val converters
   key.converter=org.apache.kafka.connect.storage.StringConverter
   value.converter=org.apache.kafka.connect.storage.StringConverter
   # http sink connector configs
   auth.type=NONE
   confluent.topic.bootstrap.servers=localhost:9092
   confluent.topic.replication.factor=1
   http.api.url=http://localhost:8080/api/messages/${topic}/${key}
   

3. Produce a set of messages with keys and values.

   Caution

   You must include a double dash (--) between the topic name and your flag. For more information, see this post.

   confluent local produce key-val-topic -- --property parse.key=true --property key.separator=,
   
   > 1,value
   > 2,another-value
   

4. Run and validate the connector as described in the Quick Start.

Tip

Run curl localhost:8080/api/messages | jq to see that the messages key and topic were saved.

Regex Replacements

The connector can be configured to match on regex.patterns and replace any matches with the regex.replacements. The regex pattern match and replacement is done after the record has been converted into its string representation.

For using multiple regex patterns, the default separator is ~ but can be configured via regex.separator.

Regex Replacement Example

1. Run the demo app with the basic-auth Spring profile.

   mvn spring-boot:run -Dspring.profiles.active=basic-auth
   

2. Create a http-sink.properties file with the following contents:

   name=RegexHttpSink
   topics=email-topic,non-email-topic
   tasks.max=1
   connector.class=io.confluent.connect.http.HttpSinkConnector
   # key/val converters
   key.converter=org.apache.kafka.connect.storage.StringConverter
   value.converter=org.apache.kafka.connect.storage.StringConverter
   # licensing for local single-node Kafka cluster
   confluent.topic.bootstrap.servers=localhost:9092
   confluent.topic.replication.factor=1
   # http sink connector configs
   http.api.url=http://localhost:8080/api/messages
   auth.type=BASIC
   connection.user=admin
   connection.password=password
   # regex to mask emails
   regex.patterns=^.+@.+$
   regex.replacements=********
   

3. Publish messages to the topics that are configured. Emails should be redacted with ******** before being sent to the demo app.

   confluent local produce email-topic
   > example@domain.com
   > another@email.com
   
   confluent local produce non-email-topic
   > not an email
   > another normal string
   

4. Run and validate the connector as described in the Quick Start.

Tombstone Records

A record that has a non-null key and a null value is refered to as a tombstone in Kafka. These records are handled specially by the HTTP Sink connector.

By default, tombstone records are ignored but this behavior can be configured with the behavior.on.null.values property.

The other two configuration options are:

• fail: If a tombstone record is received, the connector task is killed immediately.
• delete: The connector attempts to send a DELETE request to the configured API.

If key substitution is being used (ex. localhost:8080/api/messages/${key}), a DELETE request is sent to the configured URL with the key injected into the ${key} placeholder. If key substitution is not configured, the record key is appended to the end of the URI and a DELETE is sent to the formatted URL.

Delete URL Example

# EXAMPLE - KEY SUBSTITUTION

http.api.url=http://localhost:8080/api/messages/${key}
behavior.on.null.values=delete

# SinkRecord with key = 12, value = "mark@email.com"
# DELETE sent to http://localhost:8080/api/messages/12


# EXAMPLE - KEY APPENDED TO END

http.api.url=http://localhost:8080/api/messages
behavior.on.null.values=delete

# SinkRecord with key = 25, value = "jane@email.com"
# DELETE sent to http://localhost:8080/api/messages/25

Delete Behavior on Null Values Example

1. Run the demo app with the simple-auth Spring profile.

   mvn spring-boot:run -Dspring.profiles.active=basic-auth
   

2. Create a http-sink.properties file with the following contents:

   name=DeleteNullHttpSink
   topics=http-messages
   tasks.max=1
   connector.class=io.confluent.connect.http.HttpSinkConnector
   # key/val converters
   key.converter=org.apache.kafka.connect.storage.StringConverter
   value.converter=org.apache.kafka.connect.storage.StringConverter
   # licensing for local single-node Kafka cluster
   confluent.topic.bootstrap.servers=localhost:9092
   confluent.topic.replication.factor=1
   # http sink connector configs
   http.api.url=http://localhost:8080/api/messages
   auth.type=BASIC
   connection.user=admin
   connection.password=password
   behavior.on.null.values=delete
   

3. Publish messages to the topic that have keys and values.

   Caution

   You must include a double dash (--) between the topic name and your flag. For more information, see this post.

   confluent local produce http-messages -- --property parse.key=true --property key.separator=,
   > 1,message-value
   > 2,another-message
   

4. Run and validate the connector as described in the Quick Start.

   Tip

   Check for messages in the demo API with this command: curl http://localhost:8080/api/messages -H 'Authorization: Basic YWRtaW46cGFzc3dvcmQ=' | jq

5. Publish messages to the topic that have keys with null values (tombstones).

   Note

   This cannot be done with confluent local produce but there is an API in the demo app to send tombstones.

   curl -X POST \
     'localhost:8080/api/tombstone?topic=http-messages&key=1' \
     -H 'Authorization: Basic YWRtaW46cGFzc3dvcmQ='
   

6. Validate that the demo app deleted the messages.

   curl http://localhost:8080/api/messages \
     -H 'Authorization: Basic YWRtaW46cGFzc3dvcmQ=' | jq
   

Retries

In case of failures, the connector can be configured to retry the operations by configuring max.retries and retry.backoff.ms parameters.

Retries Example

1. Run the demo app with the basic-auth Spring profile.

   mvn spring-boot:run -Dspring.profiles.active=basic-auth
   

2. Create a http-sink.properties file with the following contents:

   name=RetriesExample
   topics=http-messages
   tasks.max=1
   connector.class=io.confluent.connect.http.HttpSinkConnector
   # key/val converters
   key.converter=org.apache.kafka.connect.storage.StringConverter
   value.converter=org.apache.kafka.connect.storage.StringConverter
   # licensing for local single-node Kafka cluster
   confluent.topic.bootstrap.servers=localhost:9092
   confluent.topic.replication.factor=1
   # http sink connector configs
   http.api.url=http://localhost:8080/api/messages
   auth.type=BASIC
   connection.user=admin
   connection.password=password
   behavior.on.null.values=delete
   # retry configurations
   max.retries=20
   retry.backoff.ms=5000
   

3. Publish messages to the topic that have keys and values.

   Caution

   You must include a double dash (--) between the topic name and your flag. For more information, see this post.

   confluent local produce http-messages -- --property parse.key=true --property key.separator=,
   > 1,message-value
   > 2,another-message
   

4. Stop the demo app.

5. Run and validate the connector as described in the Quick Start.

6. The Connector will retry for maximum 20 times with an initial backoff duration of 5000ms. If the http operation is successful then the retry will be stopped. In this case the connector will retry for 20 times and the connector task will get failed.

7. The default value for max.retries is 10 and for retry.backoff.ms is 3000ms.

Reporter

The Connector can be configured to capture the success/failure responses from http operations by configuring reporter parameters.

Reporter Example

1. Run the demo app with the basic-auth Spring profile.

   mvn spring-boot:run -Dspring.profiles.active=basic-auth
   

2. Create a http-sink.properties file with the following contents:

   name=ReporterExample
   topics=http-messages
   tasks.max=1
   connector.class=io.confluent.connect.http.HttpSinkConnector
   # key/val converters
   key.converter=org.apache.kafka.connect.storage.StringConverter
   value.converter=org.apache.kafka.connect.storage.StringConverter
   # licensing for local single-node Kafka cluster
   confluent.topic.bootstrap.servers=localhost:9092
   confluent.topic.replication.factor=1
   # http sink connector configs
   http.api.url=http://localhost:8080/api/messages
   auth.type=BASIC
   connection.user=admin
   connection.password=password
   behavior.on.null.values=delete
   # reporter configurations
   reporter.bootstrap.servers=localhost:9092
   reporter.result.topic.name=success-responses
   reporter.result.topic.replication.factor=1
   reporter.error.topic.name=error-responses
   reporter.error.topic.replication.factor=1
   

3. Publish messages to the topic that have keys and values.

   Caution

   You must include a double dash (--) between the topic name and your flag. For more information, see this post.

   confluent local produce http-messages -- --property parse.key=true --property key.separator=,
   > 1,message-value
   > 2,another-message
   

4. Run and validate the connector as described in the Quick Start.

5. Consume the records from success-responses and error-responses topic to see the http operation response.

   kafkacat -C -b localhost:9092 -t success-responses -J |jq
   

   {
     "topic": "success-responses",
     "partition": 0,
     "offset": 0,
     "tstype": "create",
     "ts": 1581579911854,
     "headers": [
       "input_record_offset",
       "0",
       "input_record_timestamp",
       "1581488456476",
       "input_record_partition",
       "0",
       "input_record_topic",
       "http-connect"
     ],
     "key": null,
     "payload": "{\"id\":1,\"message\":\"1,message-value\"}"
   }
   {
     "topic": "success-responses",
     "partition": 0,
     "offset": 1,
     "tstype": "create",
     "ts": 1581579911854,
     "headers": [
       "input_record_offset",
       "1",
       "input_record_timestamp",
       "1581488456476",
       "input_record_partition",
       "0",
       "input_record_topic",
       "http-connect"
     ],
     "key": null,
     "payload": "{\"id\":2,\"message\":\"2,message-value\"}"
   }
   

   In case of retryable errors (i.e., errors with a 5xx status code), a response like the one shown below is included in the error-responses topic.

   kafkacat -C -b localhost:9092 -t error-responses -J |jq
   

   {
     "topic": "error-responses",
     "partition": 0,
     "offset": 0,
     "tstype": "create",
     "ts": 1581579911854,
     "headers": [
       "input_record_offset",
       "0",
       "input_record_timestamp",
       "1581579931450",
       "input_record_partition",
       "0",
       "input_record_topic",
       "http-messages"
     ],
     "key": null,
     "payload": "Retry time lapsed, unable to process HTTP request. Error while processing HTTP request with Url : http://localhost:8080/api/messages, Payload : 6,test, Status code : 500, Reason Phrase : , Response Content : {\"timestamp\":\"2020-02-11T10:44:41.574+0000\",\"status\":500,\"error\":\"Internal Server Error\",\"message\":\"Unresolved compilation problem: \\n\\tlog cannot be resolved\\n\",\"path\":\"/api/messages\"}, "
   }
   

Additional Documentation

• Configuration Properties
• Changelog