kfc producer [--brokers=<brks>] [--partition=<part>] [--compression=<comp>]
             [--delimiter=<delim>] [--key-delimiter=<delim>] [--count=<cnt>]
             [--error-file=<file>] [-T | --tee] [-q | --quiet]
             [ (-X <conf>)... ] [-v | --verbose] <topic> [<file>...]

kfc consumer [--brokers=<brks>] [--partition=<part>] [--offset=<off>]
             [--delimiter=<delim>] [--key-delimiter=<delim>] [--count=<cnt>]
             [-e | --exit] [-O | --print-offset] [-u | --unbuffered]
             [ (-X <conf>)... ] [-q | --quiet] [-v | --verbose] <topic>

kfc metadata [--brokers=<brks>] [--partition=<part>] [-q | --quiet]
             [ (-X <conf>)... ] [-v | --verbose] [<topic>]

kfc --help

kfc --version

kfc also known as kafkacat is a generic non-JVM producer and consumer for Apache Kafka 0.8, think of it as a netcat for Kafka.

kfc works in different : producer, consumer and metadata. Each one is described in the [COMMAND] section.

• producer reads messages from stdin, delimited with a configurable delimeter and produces them to the provided Kafka cluster, topic and partition.

• consumer reads messages from a topic (and possibly a specific partition ) and prints them to stdout using the configured message delimiter.

• metadata displays the current state of the Kafka cluster and its topics and partitions.

• -b <brokers>, --brokers=<brokers> Comma separated list of broker(s) to bootstrap the connection, e.g. "host1:9292,host2:9293" [default: localhost].

• -p <partition>, --partition=<partition> Send messages to a specific partition. If -1 is provided, a random partition is attributed to each message.

• -X <conf> Relay <conf> to librdkafka configuration system. This option hides a lot of functionnality. For more information, the user should read the matching CONFIGURATION.md of librdkafka.

• -v, --verbose Augment verbosity level.

• -q, --quiet Quiet mode, do not report error messages to stderr.

• -V, --version Show the version on the first line and follow with the usage and help message.

• -h, --help Show the usage and help message.

• -d <delim>, --delimiter=<delim> Message delimiter character: a-z.. | \r | \n | \t | \xNN [default: \n].

• -k <delim>, --key-delimiter=<delim> Key delimiter character: a-z.. | \r | \n | \t | \xNN.

• -c <cnt>, --count=<cnt> Exit after producing <cnt> messages.

• -z <c>, --compression=<c> Compression applied to messages: none, snappy or gzip [default: none].

• -T, --tee Output queued messages to stdout, acting like tee. A message is relayed to stdout once it hits librdkafka's internal queue, it can still fail to reach the broker. See -E for more information on failed messages.

• -E <file>, --error-file=<file> Messages that couldn't be sent are written to <file>. There is no guarantee on the ordering of messages [default: stderr].

• -d <delim>, --delimiter=<delim> Message delimiter character: a-z.. | \r | \n | \t | \xNN [default: \n].

• -k <delim>, --key-delimiter=<delim> Key delimiter character: a-z.. | \r | \n | \t | \xNN.

• -o <off>, --offset=<off> Define the consuming offset, the possible values are:

  • beginning: start from the beginning
  • end: start from the end
  • stored: start from a stored value
  • <value>: start at absolute
  • -<value>: start at relative offset from the end (acting like tail).

• -c <cnt>, --count=<cnt> Exit after consuming <cnt> messages.

• -e , --exit Exit when there is no message left to consumer in the topic.

• -u , --unbuffered Do not buffer stdout, useful for low volume topic.

• Producing a single message

    $ echo "test message" | kfc producer test

• Consuming the last message and exiting immediately

    $ kfc consumer -e -o -1 test
    test message

• Showing metadata on the test topic

    $ kfc metadata test
    {
      "brokers": [
        { "id":0, "host":"localhost:9092" }
      ],
      "topics": [
        {
          "name": "test",
          "partitions": [
            { "id": 0, "leader": 0, "replicas": [0], "isrs": [0] }
          ]
        }
      ]
    }