[
https://issues.apache.org/jira/browse/BEAM-9439?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17070983#comment-17070983
]
Sebastian Graca commented on BEAM-9439:
---------------------------------------
We're currently testing a fix for the issue and we'll submit a patch when we're
confident that it works well in our production systems. It's just removing
{{getTotalBacklogBytes}} and using current implementation as
{{getSplitBacklogBytes}} as [~m.juraszek] suggested.
This might lead to reporting the backlog for the split that is larger than the
real one but will never report smaller backlog that would otherwise not trigger
autoscaling of Dataflow job.
> KinesisReader does not report correct backlog statistics
> ---------------------------------------------------------
>
> Key: BEAM-9439
> URL: https://issues.apache.org/jira/browse/BEAM-9439
> Project: Beam
> Issue Type: Bug
> Components: io-java-kinesis
> Reporter: Sam Whittle
> Priority: Major
>
> The KinesisReader implementing KinesisIO reports backlog by implementing the
> UnboundedSource.getTotalBacklogBytes()
> method as opposed to the
> UnboundedSource.getSplitBacklogBytes()
> This value is supposed to represent the total backlog across all shards.
> This function is implemented by calling
> SimplifiedKinesisClient.getBacklogBytes with the watermark of the kinesis
> shards managed within the UnboundedReader instance. As this watermark may be
> further ahead than the watermark across all shards, this may miss backlog
> bytes.
> An additional concern is that the watermark is calculated using a
> WatermarkPolicy, which means that the watermark may be inconsistent to the
> kinesis timestamp for querying backlog.
--
This message was sent by Atlassian Jira
(v8.3.4#803005)
