Kuan Po Tseng created KAFKA-19813:
-------------------------------------
Summary: Incorrect jitter value in
StreamsGroupHeartbeatRequestManager and AbstractHeartbeatRequestManager
Key: KAFKA-19813
URL: https://issues.apache.org/jira/browse/KAFKA-19813
Project: Kafka
Issue Type: Bug
Components: clients, consumer
Reporter: Kuan Po Tseng
Assignee: Kuan Po Tseng
see https://github.com/apache/kafka/pull/14873#discussion_r2442794887,
In below code snippets
https://github.com/apache/kafka/blob/trunk/clients/src/main/java/org/apache/kafka/clients/consumer/internals/AbstractHeartbeatRequestManager.java#L116
https://github.com/apache/kafka/blob/trunk/clients/src/main/java/org/apache/kafka/clients/consumer/internals/StreamsGroupHeartbeatRequestManager.java#L333
we use max.poll.interval.ms (default 300000) as the jitter value. Although
RequestState.remainingBackoffMs guards with Math.max(0, …), which won’t make
the value become negative, using a jitter that isn’t in (0–1) is unexpected.
In addition, we should validate that ExponentialBackoff receives a jitter
strictly within (0, 1) to prevent this scenario in the future.
--
This message was sent by Atlassian Jira
(v8.20.10#820010)
