[
https://issues.apache.org/jira/browse/KAFKA-3418?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]
Jason Gustafson updated KAFKA-3418:
-----------------------------------
Description: There still seems to be a lot of confusion about the design of
the poll() loop in regard to consumer liveness. We do mention it in the
javadoc, but it's a little hidden and we aren't very clear on what the user
should do to limit the potential for the consumer to fall out of the group
(such as tweaking max.poll.records). We should pull this into a separate
section (e.g. Jay suggests "Detecting Consumer Failures") and give it a more
complete treatment. (was: There still seems to be a lot of confusion about the
design of the poll() loop in regard to consumer liveness. We do mention it in
the javadoc, but it's a little hidden and we aren't very clear on what the user
should do to (such as tweaking max.poll.records). We should pull this into a
separate section (e.g. Jay suggests "Detecting Consumer Failures") and give it
a more complete treatment.)
> Add section on detecting consumer failures in new consumer javadoc
> ------------------------------------------------------------------
>
> Key: KAFKA-3418
> URL: https://issues.apache.org/jira/browse/KAFKA-3418
> Project: Kafka
> Issue Type: Improvement
> Components: consumer
> Reporter: Jason Gustafson
> Assignee: Jason Gustafson
> Fix For: 0.10.0.0
>
>
> There still seems to be a lot of confusion about the design of the poll()
> loop in regard to consumer liveness. We do mention it in the javadoc, but
> it's a little hidden and we aren't very clear on what the user should do to
> limit the potential for the consumer to fall out of the group (such as
> tweaking max.poll.records). We should pull this into a separate section (e.g.
> Jay suggests "Detecting Consumer Failures") and give it a more complete
> treatment.
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
