[
https://issues.apache.org/jira/browse/FLINK-2292?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14626065#comment-14626065
]
ASF GitHub Bot commented on FLINK-2292:
---------------------------------------
Github user mxm commented on the pull request:
https://github.com/apache/flink/pull/896#issuecomment-121173161
> Wasn't the rationale to have a separate message for the heartbeat and the
accumulators to keep the heartbeat messages small? What do you think? In any
case, I agree that it makes sense to be able to configure this.
Initially, that was my proposal but I remember we discussed that we should
try to minimize the amount of messages and thus use the `Heartbeat` message. It
makes sense to be able to configure the snapshot transferal in multiples of the
hearbeat interval. I'd like to fix that in another pull
> I would add a unit test for the record reader/writer accumulator results.
I'll integrate that into the `AccumulatorLiveITCase`.
> There seems to be a 1:1 correspondence between the reporter and the
internal metrics. Is the idea to have multiple different reporters in the
future (for different types of metrics etc.) or is a long reporter sufficient?
For now, a long reporter is sufficient but if we discover that we need more
customizable reporters, we can easily change the reporter interface in the
future (it wouldn't be API breaking).
> Report accumulators periodically while job is running
> -----------------------------------------------------
>
> Key: FLINK-2292
> URL: https://issues.apache.org/jira/browse/FLINK-2292
> Project: Flink
> Issue Type: Sub-task
> Components: JobManager, TaskManager
> Reporter: Maximilian Michels
> Assignee: Maximilian Michels
> Fix For: 0.10
>
>
> Accumulators should be sent periodically, as part of the heartbeat that sends
> metrics. This allows them to be updated in real time.
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
