[
https://issues.apache.org/jira/browse/FLINK-34400?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17815218#comment-17815218
]
Rui Fan commented on FLINK-34400:
---------------------------------
{quote} One topic remains pretty much empty most of the time, while the other
receives a few events per second all the time.
I think that if I add {{.withIdleness(...)}} to the watermark strategy, it also
works fine, but I will do some more testing.
{quote}
After re-read the background, I guess it may be not a bug.
When one source or topic doesn't have data, the watermark of this source won't
be update. And then watermark alignment will let other quick sources to wait
for this slow source.
As you said, I think idleness is suitable for your case. WDYT?
> Kafka sources with watermark alignment sporadically stop consuming
> ------------------------------------------------------------------
>
> Key: FLINK-34400
> URL: https://issues.apache.org/jira/browse/FLINK-34400
> Project: Flink
> Issue Type: Bug
> Affects Versions: 1.18.1
> Reporter: Alexis Sarda-Espinosa
> Priority: Major
> Attachments: alignment_lags.png, logs.txt
>
>
> I have 2 Kafka sources that read from different topics. I have assigned them
> to the same watermark alignment group, and I have _not_ enabled idleness
> explicitly in their watermark strategies. One topic remains pretty much empty
> most of the time, while the other receives a few events per second all the
> time. Parallelism of the active source is 2, for the other one it's 1, and
> checkpoints are once every minute.
> This works correctly for some time (10 - 15 minutes in my case) but then 1 of
> the active sources stops consuming, which causes lag to increase. Weirdly,
> after another 15 minutes or so, all the backlog is consumed at once, and then
> everything stops again.
> I'm attaching some logs from the Task Manager where the issue appears. You
> will notice that the Kafka network client reports disconnections (a long time
> after the deserializer stopped reporting that events were being consumed),
> I'm not sure if this is related.
--
This message was sent by Atlassian Jira
(v8.20.10#820010)
