[
https://issues.apache.org/jira/browse/FLINK-1297?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14499809#comment-14499809
]
ASF GitHub Bot commented on FLINK-1297:
---------------------------------------
Github user StephanEwen commented on the pull request:
https://github.com/apache/flink/pull/605#issuecomment-93984865
I think this is nice work.
It would be easier to add and maintain this, if it were not hardwired into
the core flink classes, but if we could make this like a library. This would be
a good example for the `flink-contrib` project.
The changes for this should be rather small: It would boil down to not
adding the specialized methods for the `OperatorStatistics` to the
RuntimeEnvironment and the `JobExecutionResult`, but to add a util class that
converts an operator statistics accumulator into the OperatorStatistics object.
I general, I am a bit more comfortable with adding stuff as libraries than
to the core API. A lot of requests and features are coming in, if we all add
them to the core API, it will easily become unmaintainable. Staging something
first as a library is pretty uncritical. We can always add it to the API later,
if it becomes stable and heavily used.
> Add support for tracking statistics of intermediate results
> -----------------------------------------------------------
>
> Key: FLINK-1297
> URL: https://issues.apache.org/jira/browse/FLINK-1297
> Project: Flink
> Issue Type: Improvement
> Components: Distributed Runtime
> Reporter: Alexander Alexandrov
> Assignee: Alexander Alexandrov
> Fix For: 0.9
>
> Original Estimate: 1,008h
> Remaining Estimate: 1,008h
>
> One of the major problems related to the optimizer at the moment is the lack
> of proper statistics.
> With the introduction of staged execution, it is possible to instrument the
> runtime code with a statistics facility that collects the required
> information for optimizing the next execution stage.
> I would therefore like to contribute code that can be used to gather basic
> statistics for the (intermediate) result of dataflows (e.g. min, max, count,
> count distinct) and make them available to the job manager.
> Before I start, I would like to hear some feedback form the other users.
> In particular, to handle skew (e.g. on grouping) it might be good to have
> some sort of detailed sketch about the key distribution of an intermediate
> result. I am not sure whether a simple histogram is the most effective way to
> go. Maybe somebody would propose another lightweight sketch that provides
> better accuracy.
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
