Module: Karafka::Connection::BatchDelegator

Defined in:
lib/karafka/connection/batch_delegator.rb

Overview

Class that delegates processing of batch received messages for which we listen to a proper processor

Class Method Summary collapse

Class Method Details

.call(group_id, kafka_batch) ⇒ Object

Note:

This should be looped to obtain a constant delegating of new messages

Delegates messages (does something with them) It will either schedule or run a proper processor action for messages

Parameters:

  • group_id (String) —

    group_id of a group from which a given message came

  • kafka_batch (<Kafka::FetchedBatch>) —

    raw messages fetched batch



14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
 
# File 'lib/karafka/connection/batch_delegator.rb', line 14

def call(group_id, kafka_batch)
  topic = Persistence::Topics.fetch(group_id, kafka_batch.topic)
  consumer = Persistence::Consumers.fetch(topic, kafka_batch.partition)

  Karafka.monitor.instrument(
    'connection.batch_delegator.call',
    caller: self,
    consumer: consumer,
    kafka_batch: kafka_batch
  ) do
    # Due to how ruby-kafka is built, we have the metadata that is stored on the batch
    # level only available for batch consuming
    consumer.batch_metadata = Params::Builders::BatchMetadata.from_kafka_batch(
      kafka_batch,
      topic
    )

    kafka_messages = kafka_batch.messages

    # Depending on a case (persisted or not) we might use new consumer instance per
    # each batch, or use the same one for all of them (for implementing buffering, etc.)
    if topic.batch_consuming
      consumer.params_batch = Params::Builders::ParamsBatch.from_kafka_messages(
        kafka_messages,
        topic
      )
      consumer.call
    else
      kafka_messages.each do |kafka_message|
        consumer.params_batch = Params::Builders::ParamsBatch.from_kafka_messages(
          [kafka_message],
          topic
        )
        consumer.call
      end
    end
  end
end