moshe blumberg created KAFKA-7073:
-------------------------------------
Summary: Allow retries to try different partitions where the use
case allows it
Key: KAFKA-7073
URL: https://issues.apache.org/jira/browse/KAFKA-7073
Project: Kafka
Issue Type: Improvement
Affects Versions: 1.0.0
Reporter: moshe blumberg
Currently producer retry logic will always retry to the same partition.
However, this means that if that partition leader is slow or unavailable the
retry will suffer.
In use cases where messages are not sent to a specific partition this
limitation is unnecessary. The retries could send to another, performing,
partition.
Could an indication of whether specific partitioning is required and the retry
logic behave accordingly?
--
This message was sent by Atlassian JIRA
(v7.6.3#76005)
