Kirk True created KAFKA-18571:
---------------------------------
Summary: Investigate the use of metadata caching to improve fetch
request generation performance
Key: KAFKA-18571
URL: https://issues.apache.org/jira/browse/KAFKA-18571
Project: Kafka
Issue Type: Improvement
Components: clients, consumer
Affects Versions: 4.0.0
Reporter: Kirk True
Assignee: Kirk True
During a [review|https://github.com/apache/kafka/pull/17700] of the design and
implementation of KAFKA-17182, a [question arose around the CPU performance
impact|https://github.com/apache/kafka/pull/17700#discussion_r1915249053]. In
the ensuing discussion, the idea was put forth that pre-processing and/or
caching some of Node, TopicPartition, and FetchBuffer data used during the
fetch process could improve performance.
cc [~jeffkbkim]
--
This message was sent by Atlassian Jira
(v8.20.10#820010)
