Kafdrop – Kafka Web UI  

Kafdrop is a web UI for viewing Kafka topics and browsing consumer groups. The tool displays information such as brokers, topics, partitions, consumers, and lets you view messages.

This project is a reboot of Kafdrop 2.x, dragged kicking and screaming into the world of Java 17+, Kafka 2.x, Helm and Kubernetes. It's a lightweight application that runs on Spring Boot and is dead-easy to configure, supporting SASL and TLS-secured brokers.

Features

• View Kafka brokers — topic and partition assignments, and controller status
• View topics — partition count, replication status, and custom configuration
• Browse messages — JSON, plain text, Avro and Protobuf encoding
• View consumer groups — per-partition parked offsets, combined and per-partition lag
• Create new topics
• View ACLs
• Support for Azure Event Hubs

Requirements

• Java 17 or newer
• Kafka (version 0.11.0 or newer) or Azure Event Hubs

Optional, additional integration:

• Schema Registry

Getting Started

You can run the Kafdrop JAR directly, via Docker, or in Kubernetes.

Running from JAR

java --add-opens=java.base/sun.nio.ch=ALL-UNNAMED \
    -jar target/kafdrop-<version>.jar \
    --kafka.brokerConnect=<host:port,host:port>,...

If unspecified, kafka.brokerConnect defaults to localhost:9092.

Note: As of Kafdrop 3.10.0, a ZooKeeper connection is no longer required. All necessary cluster information is retrieved via the Kafka admin API.

Open a browser and navigate to http://localhost:9000. The port can be overridden by adding the following config:

--server.port=<port> --management.server.port=<port>

Optionally, configure a schema registry connection with:

--schemaregistry.connect=http://localhost:8081

and if you also require basic auth for your schema registry connection you should add:

--schemaregistry.auth=username:password

Finally, a default message and key format (e.g. to deserialize Avro messages or keys) can optionally be configured as follows:

--message.format=AVRO
--message.keyFormat=DEFAULT

Valid format values are DEFAULT, AVRO, PROTOBUF. This can also be configured at the topic level via dropdown when viewing messages. If key format is unspecified, message format will be used for key too.

Configure Protobuf message type

Option 1: Using Protobuf Descriptor

In case of protobuf message type, the definition of a message could be compiled and transmitted using a descriptor file. Thus, in order for kafdrop to recognize the message, the application will need to access to the descriptor file(s). Kafdrop will allow user to select descriptor and well as specifying name of one of the message type provided by the descriptor at runtime.

To configure a folder with protobuf descriptor file(s) (.desc), follow:

--protobufdesc.directory=/var/protobuf_desc

Option 2 : Using Schema Registry

In case of no protobuf descriptor file being supplied the implementation will attempt to create the protobuf deserializer using the schema registry instead.

Defaulting to Protobuf

If preferred the message type could be set to default as follows:

--message.format=PROTOBUF

Running with Docker

Images are hosted at hub.docker.com/r/obsidiandynamics/kafdrop.

Launch container in background:

docker run -d --rm -p 9000:9000 \
    -e KAFKA_BROKERCONNECT=<host:port,host:port> \
    -e SERVER_SERVLET_CONTEXTPATH="/" \
    obsidiandynamics/kafdrop

Launch container with some specific JVM options:

docker run -d --rm -p 9000:9000 \
    -e KAFKA_BROKERCONNECT=<host:port,host:port> \
    -e JVM_OPTS="-Xms32M -Xmx64M" \
    -e SERVER_SERVLET_CONTEXTPATH="/" \
    obsidiandynamics/kafdrop

Launch container in background with protobuff definitions:

docker run -d --rm -v <path_to_protobuff_descriptor_files>:/var/protobuf_desc -p 9000:9000 \
    -e KAFKA_BROKERCONNECT=<host:port,host:port> \
    -e SERVER_SERVLET_CONTEXTPATH="/" \
    -e CMD_ARGS="--message.format=PROTOBUF --protobufdesc.directory=/var/protobuf_desc" \
    obsidiandynamics/kafdrop

Then access the web UI at http://localhost:9000.

Hey there! We hope you really like Kafdrop! Please take a moment to ⭐the repo or Tweet about it.

Running in Kubernetes (using a Helm Chart)

Clone the repository (if necessary):

git clone https://github.com/obsidiandynamics/kafdrop && cd kafdrop

Apply the chart:

helm upgrade -i kafdrop chart --set image.tag=3.x.x \
    --set kafka.brokerConnect=<host:port,host:port> \
    --set server.servlet.contextPath="/" \
    --set cmdArgs="--message.format=AVRO --schemaregistry.connect=http://localhost:8080" \ #optional
    --set jvm.opts="-Xms32M -Xmx64M"

For all Helm configuration options, have a peek into chart/values.yaml.

Replace 3.x.x with the image tag of obsidiandynamics/kafdrop. Services will be bound on port 9000 by default (node port 30900).

Note: The context path must begin with a slash.

Proxy to the Kubernetes cluster:

kubectl proxy

Navigate to http://localhost:8001/api/v1/namespaces/default/services/http:kafdrop:9000/proxy.

Protobuf support via helm chart:

To install with protobuf support, a "facility" option is provided for the deployment, to mount the descriptor files folder, as well as passing the required CMD arguments, via option mountProtoDesc. Example:

helm upgrade -i kafdrop chart --set image.tag=3.x.x \
    --set kafka.brokerConnect=<host:port,host:port> \
    --set server.servlet.contextPath="/" \
    --set mountProtoDesc.enabled=true \
    --set mountProtoDesc.hostPath="<path/to/desc/folder>" \
    --set jvm.opts="-Xms32M -Xmx64M"

Building

After cloning the repository, building is just a matter of running a standard Maven build:

$ mvn clean package

The following command will generate a Docker image:

mvn assembly:single docker:build

Docker Compose

There is a docker-compose.yaml file that bundles a Kafka/ZooKeeper instance with Kafdrop:

cd docker-compose/kafka-kafdrop
docker-compose up

APIs

JSON endpoints

Starting with version 2.0.0, Kafdrop offers a set of Kafka APIs that mirror the existing HTML views. Any existing endpoint can be returned as JSON by simply setting the Accept: application/json header. Some endpoints are JSON only:

• /topic: Returns a list of all topics.

OpenAPI Specification (OAS)

To help document the Kafka APIs, OpenAPI Specification (OAS) has been included. The OpenAPI Specification output is available by default at the following Kafdrop URL:

/v3/api-docs

It is also possible to access the Swagger UI (the HTML views) from the following URL:

/swagger-ui.html

This can be overridden with the following configuration:

springdoc.api-docs.path=/new/oas/path

You can disable OpenAPI Specification output with the following configuration:

springdoc.api-docs.enabled=false

CORS Headers

Starting in version 2.0.0, Kafdrop sets CORS headers for all endpoints. You can control the CORS header values with the following configurations:

cors.allowOrigins (default is *)
cors.allowMethods (default is GET,POST,PUT,DELETE)
cors.maxAge (default is 3600)
cors.allowCredentials (default is true)
cors.allowHeaders (default is Origin,Accept,X-Requested-With,Content-Type,Access-Control-Request-Method,Access-Control-Request-Headers,Authorization)

You can also disable CORS entirely with the following configuration:

cors.enabled=false

Topic Configuration

By default, you could delete a topic. If you don't want this feature, you could disable it with:

--topic.deleteEnabled=false

By default, you could create a topic. If you don't want this feature, you could disable it with:

--topic.createEnabled=false

Actuator

Health and info endpoints are available at the following path: /actuator

This can be overridden with the following configuration:

management.endpoints.web.base-path=<path>

Guides

Connecting to a Secure Broker

Kafdrop supports TLS (SSL) and SASL connections for encryption and authentication. This can be configured by providing a combination of the following files (placed into the Kafka root directory):

• kafka.truststore.jks: specifying the certificate for authenticating brokers, if TLS is enabled.
• kafka.keystore.jks: specifying the private key to authenticate the client to the broker, if mutual TLS authentication is required.
• kafka.properties: specifying the necessary configuration, including key/truststore passwords, cipher suites, enabled TLS protocol versions, username/password pairs, etc. When supplying the truststore and/or keystore files, the ssl.truststore.location and ssl.keystore.location properties will be assigned automatically.

Using Docker

The three files above can be supplied to a Docker instance in base-64-encoded form via environment variables:

docker run -d --rm -p 9000:9000 \
    -e KAFKA_BROKERCONNECT=<host:port,host:port> \
    -e KAFKA_PROPERTIES="$(cat kafka.properties | base64)" \
    -e KAFKA_TRUSTSTORE="$(cat kafka.truststore.jks | base64)" \   # optional
    -e KAFKA_KEYSTORE="$(cat kafka.keystore.jks | base64)" \       # optional
    obsidiandynamics/kafdrop

Rather than passing KAFKA_PROPERTIES as a base64-encoded string, you can also place a pre-populated KAFKA_PROPERTIES_FILE into the container:

cat << EOF > kafka.properties
security.protocol=SASL_SSL
sasl.mechanism=SCRAM-SHA-512
sasl.jaas.config=org.apache.kafka.common.security.scram.ScramLoginModule required username="foo" password="bar"
EOF

docker run -d --rm -p 9000:9000 \
    -v $(pwd)/kafka.properties:/tmp/kafka.properties:ro \
    -v $(pwd)/kafka.truststore.jks:/tmp/kafka.truststore.jks:ro \
    -v $(pwd)/kafka.keystore.jks:/tmp/kafka.keystore.jks:ro \
    -e KAFKA_BROKERCONNECT=<host:port,host:port> \
    -e KAFKA_PROPERTIES_FILE=/tmp/kafka.properties \
    -e KAFKA_TRUSTSTORE_FILE=/tmp/kafka.truststore.jks \   # optional
    -e KAFKA_KEYSTORE_FILE=/tmp/kafka.keystore.jks \       # optional
    obsidiandynamics/kafdrop

It's sometimes needed to load extra classes, e.g. for a SASL client callback handler. To facilitate that, it is possible to mount a folder with extra JARs, like this:

cat << EOF > kafka.properties
security.protocol=SASL_SSL
sasl.jaas.config=software.amazon.msk.auth.iam.IAMLoginModule;
sasl.client.callback.handler.class=software.amazon.msk.auth.iam.IAMClientCallbackHandler
EOF

mkdir extra-kafdrop-classes
wget --directory-prefix=extra-kafdrop-classes https://repo1.maven.org/maven2/software/amazon/msk/aws-msk-iam-auth/1.0.0/aws-msk-iam-auth-1.0.0.jar

docker run -d --rm -p 9000:9000 \
    -v $(pwd)/kafka.properties:/tmp/kafka.properties:ro \
    -v $(pwd)/kafka.truststore.jks:/tmp/kafka.truststore.jks:ro \
    -v $(pwd)/kafka.keystore.jks:/tmp/kafka.keystore.jks:ro \
    -v $(pwd)/extra-kafdrop-classes:/extra-classes:ro \
    -e KAFKA_BROKERCONNECT=<host:port,host:port> \
    -e KAFKA_PROPERTIES_FILE=/tmp/kafka.properties \
    -e KAFKA_TRUSTSTORE_FILE=/tmp/kafka.truststore.jks \   # optional
    -e KAFKA_KEYSTORE_FILE=/tmp/kafka.keystore.jks \       # optional
    obsidiandynamics/kafdrop

Environment Variables

Basic configuration

Advanced configuration

Using Helm

Like in the Docker example, supply the files in base-64 form:

helm upgrade -i kafdrop chart --set image.tag=3.x.x \
    --set kafka.brokerConnect=<host:port,host:port> \
    --set kafka.properties="$(cat kafka.properties | base64)" \
    --set kafka.truststore="$(cat kafka.truststore.jks | base64)" \
    --set kafka.keystore="$(cat kafka.keystore.jks | base64)"

Updating the Bootstrap theme

Edit the .scss files in the theme directory, then run theme/install.sh. This will overwrite src/main/resources/static/css/bootstrap.min.css. Then build as usual. (Requires npm.)

Securing the Kafdrop UI

Kafdrop doesn't (yet) natively implement an authentication mechanism to restrict user access. Here's a quick workaround using NGINX using Basic Auth. The instructions below are for macOS and Homebrew.

Requirements

• NGINX: install using which nginx > /dev/null || brew install nginx
• Apache HTTP utilities: which htpasswd > /dev/null || brew install httpd

Setup

Set the admin password (you will be prompted):

htpasswd -c /usr/local/etc/nginx/.htpasswd admin

Add a logout page in /usr/local/opt/nginx/html/401.html:

<!DOCTYPE html>
<p>Not authorized. <a href="<!--# echo var="scheme" -->://<!--# echo var="http_host" -->/">Login</a>.</p>

Use the following snippet for /usr/local/etc/nginx/nginx.conf:

worker_processes 4;

events {
  worker_connections 1024;
}

http {
  upstream kafdrop {
    server 127.0.0.1:9000;
    keepalive 64;
  }

  server {
    listen *:8080;
    server_name _;
    access_log /usr/local/var/log/nginx/nginx.access.log;
    error_log /usr/local/var/log/nginx/nginx.error.log;
    auth_basic "Restricted Area";
    auth_basic_user_file /usr/local/etc/nginx/.htpasswd;

    location / {
      proxy_pass http://kafdrop;
    }

    location /logout {
      return 401;
    }

    error_page 401 /errors/401.html;

    location /errors {
      auth_basic off;
      ssi        on;
      alias /usr/local/opt/nginx/html;
    }
  }
}

Run NGINX:

nginx

Or reload its configuration if already running:

nginx -s reload

To logout, browse to /logout.

Hey there! We hope you really like Kafdrop! Please take a moment to ⭐the repo or Tweet about it.

Contributing Guidelines

See here.

Release workflow

To cut an official release, these are the steps:

1. Commit a new version on master that has the -SNAPSHOT suffix stripped (see pom.xml). Once the commit is merged, the CI will treat it as a release build, and will end up publishing more artifacts than the regular (non-release/snapshot) build. One of those will be a dockerhub push to the specific version and "latest" tags. (The regular build doesn't update "latest").

2. You can then edit the release description in GitHub to describe what went into the release.

3. After the release goes through successfully, you need to prepare the repo for the next version, which requires committing the next snapshot version on master again. So we should increment the minor version and add again the -SNAPSHOT suffix.