Apache Kafka to WebSocket creates web-socket server that serves data from Apache Kafka topic(s). This is also great tool for inspecting Kafka topic(s) using built-in browser based consumer.

How to use

• create your config.yaml file, you can call k2ws -init to begin with
• download and run binary for your OS (config.yaml should be in the same directory)

Simple setup

schema.version: "1.0"
kafka.to.websocket:
  - kafka.consumer.config:
      metadata.broker.list: kafka:9092
      group.id: k2ws-group
    kafka.topics:
      - test
    address: :80

Complex setup

schema.version: "1.0"
# domain certificate files are located in certs folder
tls.cert.file: ./certs/k2ws.socialgist.com.crt
tls.key.file: ./certs/k2ws.socialgist.com.key
kafka.to.websocket:
  # WebSocket X | wss://k2ws.socialgist.com/kafkaA
  - kafka.consumer.config:
      metadata.broker.list: kafkaA:9092
      group.id: k2ws-wsx-group
    kafka.topics:
      - a1
      - a2
    address: :443
    endpoint.prefix: kafka_a
  # WebSocket X | wss://k2ws.socialgist.com/kafkaB
  - kafka.consumer.config:
      metadata.broker.list: kafkaB:9092
      group.id: k2ws-wsx-group
    kafka.topics:
      - b1
    address: :443
    endpoint.prefix: kafka_b
  # WebSocket Y | wss://k2ws.socialgist.com/kafkaC
  - kafka.consumer.config:
      metadata.broker.list: kafkaC:9092
      group.id: k2ws-wsy-group
    kafka.topics:
      - c1
    address: :443
    endpoint.prefix: kafka_c

Kafka config options

For setting up kafka.consumer.config and kafka.default.topic.config refer to librdkafka docs.

Property	Required	Range	Default	Description
kafka.consumer.config/metadata.broker.list	yes			Initial list of brokers as a CSV list of broker host or host:port.
kafka.consumer.config/group.id				Client group id string. All clients sharing the same group.id belong to the same group. If omitted group id will be expected to be passed by client.
kafka.default.topic.config/auto.offset.reset		smallest, earliest, beginning, largest, latest, end, error		Action to take when there is no initial offset in offset store or the desired offset is out of range. If omitted it will be expected to be passed by client.
kafka.topics				List of Kafka topics that will be served via websocket. If omitted topic list will be expected to be passed by client.
address	yes			Host (or IP) and port pair where socket will be served from. Host (IP) is optional. Example localhost:8888 or :8888.
endpoint.prefix				Prefix of the websocket and test paths. By default is empty.
endpoint.websocket				Path to websocket URL. By default it's empty.
endpoint.test			test	Path to test page URL.
message.details			false	Include key, headers, topic partition, timestamp into websocket message payload. Message will be in JSON format. Field value will hold the original message, whose representation depends on message.type property.
message.type		json, text, binary	json	Type of Kafka messages. This is only important when message.details option is set to true because it will affect creation of websocket message payload. binary messages will be encoded into base64 string, text into JSON string and json will keep original form.

When topics, kafka.consumer.config/group.id and/or kafka.consumer.config/auto.offset.reset are omitted in configuration, they are expected to be set by client as a query parameters in websocket URL. For example if websocket URL is ws://localhost:8888/ client can set these by making request to ws://localhost:8888/?topics=topicA,topicB&group.id=mygroup&auto.offset.reset=earliest. Note that this only works for parameters that are omitted from configuration thus setting them otherwise will have no effect.

You can serve more then one Kafka config entry on the same port as long as they all have unique websocket and test endpoints.

Test page

Every kafka.to.websocket entry will have it's own test page. By default this page will be accessible on /test path or on endpoint.test path if endpoint.test is defined in entry's configuration.

Test page can open up websocket and start consuming Kafka topic(s). It will display message for 1 second by default and then it will display another message that came at the current time, meaning that some messages will not be shown when message rate is high (higher then 1 message per second). Even though messages are not displayed, they will still be consumed and processed by the page. Test page has handy filtering capability that makes it possible to execute small JavaScript code for each message in which you can decide what you want to do with the message and want you want to display.

• Open - Opens up websocket connection (kafka topic(s) consumption will start).
• Close - Closes websocket connection (kafka topic(s) consumption will stop).
• Setup - Toggle button, shows or hides connection setup options.If any of topics, group.id or auto.offset.reset are omitted from configuration entry, client will have to set them up here before opening websocket connection.
• Auto-close - Defines how many messages test page should receive before it automatically closes up websocket connection.
• Stack - Toggle button, when active it will make test page stack incoming messages into JSON array.
• Record - Toggle button, when active it will start recording messages and save them to a file once button is deactivated or if size exceeds 200MB
• 5511 / 9mps - Shows total number of received messages and current message consumption rate, it will reset when it's clicked.
• 1 sec - Configures how long should a message be displayed.
• Filter - Toggle button, when active it will show filter editor where you can configure JavaScript code to execute for each message.
• Freeze - Toggle button, when active it will freeze current message, consumption will still run in the background.
• Evaluate - It will evaluate current JavaScript filter code on the latest message (or froze message if freeze is active). Only shows up when filter is active.

Incoming (or filtered) messages are displayed on the bottom left text area. Filter editor is displayed on the bottom right when filtering is activated.

Updating static HTTP files

Add/remove/modify files in k2ws/static directory and using esc tool from inside k2ws directory run esc -o static.go static. This will re-create static.go.

Build

This project has librdkafka dependency.

You can build it statically with go build -tags static. Check out confluentinc/confluent-kafka-go for more info.

Build with Docker (Linux)

docker build -f Dockerfile.build -t k2ws-build .
docker run --rm -v $PWD/build:/build k2ws-build

You'll end up with ./build/k2ws executable that works on Linux.

Build and run with Docker

Build

docker build -f Dockerfile.k2ws -t k2ws .

Run

docker run -d --name=k2ws --net=host -v _PATH_TO_CONFIG_:/config.yaml k2ws

Build for Windows

• get and start cygwin64 installation from https://www.cygwin.com/setup-x86_64.exe
  • select install from Internet
  • select C:\cygwin64 as root directory
  • select muug.ca domain
  • view -> full
  • install x86_64-w64-mingw32-gcc and pkg-config
    • click on the icon left from Skip ( keep clicking until you figure the latest version )
  • optionally put setup-x86_64.exe in c:\cygwin64\ for convenience so that you can install more libs later if needed
• add c:\cygwin64\ to PATH
• open your project directory in command prompt
• get nuget from https://dist.nuget.org/win-x86-commandline/latest/nuget.exe
• run nuget install librdkafka.redist -Version 0.11.6 ( this is currently the latest version )
• this will download rdkafka into new librdkafka.redist.0.11.6 directory
• copy .\librdkafka.redist.0.11.6\build\native\include\ into c:\cygwin64\usr\include\
• copy .\librdkafka.redist.0.11.6\build\native\lib\win7\x64\win7-x64-Release\v120\librdkafka.lib into c:\cygwin64\lib\librdkafka.a (notice .lib is renamed to .a)
• create file rdkafka.pc in the project's root directory with following content:

prefix=c:/
libdir=c:/cygwin64/lib/
includedir=c:/cygwin64/usr/include

Name: librdkafka
Description: The Apache Kafka C/C++ library
Version: 0.11.6
Cflags: -I${includedir}
Libs: -L${libdir} -lrdkafka
Libs.private: -lssl -lcrypto -lcrypto -lz -ldl -lpthread -lrt

• run set CC=x86_64-w64-mingw32-gcc (this will allow cgo to use x86_64-w64-mingw32-gcc instead of gcc - you can make sure it worked with go env CC)
• run go build, that will create executable
• deliver librdkafka.dll, msvcr120.dll and zlib.dll from .\librdkafka.redist.0.11.6\runtimes\win7-x64\native\ alongside with executable

Build for MacOS

• install openssl with brew install openssl
• run

CFLAGS="-I/usr/local/opt/openssl/include"
LDFLAGS="-L/usr/local/opt/openssl/lib"
go build -tags static