[
https://issues.apache.org/jira/browse/FLINK-3428?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15191338#comment-15191338
]
ASF GitHub Bot commented on FLINK-3428:
---------------------------------------
Github user StephanEwen commented on the pull request:
https://github.com/apache/flink/pull/1764#issuecomment-195489648
I think this is a useful tool. Few comments, though:
- You can still have a negative underflow (wrap around, positive) in the
watermark generation.
- We have somehow reached the conclusion to use Guava as little as
possible, potentially getting rid of it in the future.
- If the `extractWatermark(T, long)` and `getCurrentWatermark()` method
should not be overridden, they should probably be final.
### About the name
I am not sure about the name, maybe a native speaker can chime in here. My
git feeling was that this "trailing" would be understood as "event time trails
by a certain amount behind processing time".
I used the term "max out of orderness" in the example in the documentation
for the same case solved here:
https://ci.apache.org/projects/flink/flink-docs-release-1.0/apis/streaming/event_timestamps_watermarks.html#timestamp-assigners--watermark-generators
> Add fixed time trailing timestamp/watermark extractor
> -----------------------------------------------------
>
> Key: FLINK-3428
> URL: https://issues.apache.org/jira/browse/FLINK-3428
> Project: Flink
> Issue Type: Improvement
> Reporter: Robert Metzger
> Assignee: Kostas Kloudas
>
> Flink currently provides only one build-in timestamp extractor, which assumes
> strictly ascending timestamps. In real world use cases, timestamps are almost
> never strictly ascending.
> Therefore, I propose to provide an utility watermark extractor which is
> generating watermarks with a fixed-time trailing.
> The implementation should keep track of the highest event-time seen so far
> and subtract a fixed amount of time from that event time.
> This way, users can for example specify that the watermarks should always
> "lag behind" 10 minutes.
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
