You are viewing a plain text version of this content. The canonical link for it is here.
Posted to issues@storm.apache.org by "Anand Hariharan (Jira)" <ji...@apache.org> on 2024/08/16 10:05:00 UTC

[jira] [Created] (STORM-4076) KafkaTridentSpoutEmitters can poll all partitions at once instead of one at a time

Anand Hariharan created STORM-4076:
--------------------------------------

             Summary: KafkaTridentSpoutEmitters can poll all partitions at once instead of one at a time
                 Key: STORM-4076
                 URL: https://issues.apache.org/jira/browse/STORM-4076
             Project: Apache Storm
          Issue Type: Improvement
          Components: storm-client, storm-kafka-client
            Reporter: Anand Hariharan


{{Currently 'KafkaTridentTransactionalSpoutEmitter' and 'KafkaTridentOpaqueEmitter' polls every partition assigned to the spout one by one while emitting new batches. }}

{{But this can be improved by leveraging Kafka Consumer's usual polling strategy. That is Kafka Broker will take care of choosing the right partition.}}

{{Advantages of this are}}
{{1. If a spout is assigned multiple Topic Partitions, the consumer doesn't have to waste time on polling partitions with no data.}}
{{2. This change will give better control over the Trident batch size, by adjusting Kafka Consumer properties.}}

{{Note: This change will affect only when the batch is emitted for the first time.}}

 



--
This message was sent by Atlassian Jira
(v8.20.10#820010)