[
https://issues.apache.org/jira/browse/FLINK-21120?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]
zlzhang0122 updated FLINK-21120:
--------------------------------
Description: When using eventTime and extract watermark from element
timestamp to generate BoundedOutOfOrderness waternark, if one of the element
timestamp is abnormal(assume the timestamp is very big, for example
Integer.MAX_VALUE), then the watermark became very big too, and then the normal
element timestamp can not trigger the window bellow any more. I think maybe
there can provide a option to just ignore the abnormal element timestamp or
generate a normal watermark base on the current processTime. (was: When using
eventTime and extract watermark from element timestamp to generate
BoundedOutOfOrderness waternark, if one of the element timestamp is
abnormal(the timestamp is very big, for example Integer.MAX_VALUE), then the
watermark became very big too, and then the normal element timestamp can not
trigger the window bellow any more. I think maybe there can provide a option to
just ignore the abnormal element timestamp or generate a normal watermark base
on the current processTime.)
> Add support for ignore abnormal watermark
> -----------------------------------------
>
> Key: FLINK-21120
> URL: https://issues.apache.org/jira/browse/FLINK-21120
> Project: Flink
> Issue Type: Improvement
> Components: API / DataStream
> Affects Versions: 1.12.0, 1.11.1
> Reporter: zlzhang0122
> Priority: Critical
>
> When using eventTime and extract watermark from element timestamp to generate
> BoundedOutOfOrderness waternark, if one of the element timestamp is
> abnormal(assume the timestamp is very big, for example Integer.MAX_VALUE),
> then the watermark became very big too, and then the normal element timestamp
> can not trigger the window bellow any more. I think maybe there can provide a
> option to just ignore the abnormal element timestamp or generate a normal
> watermark base on the current processTime.
--
This message was sent by Atlassian Jira
(v8.3.4#803005)
