[
https://issues.apache.org/jira/browse/BEAM-8314?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16939758#comment-16939758
]
Ankur Goenka commented on BEAM-8314:
------------------------------------
We are refining the PR to make make it more readable and smaller.
Also we are adding tests for the change which will give us confidence.
We tried running the pipeline overnight with this fix and found it working in
expected manner.
However, CounterUpdate structure is very generic and not very well defined
which makes it harder to be 100% confident.
Given the size and number of work items in streaming, we will anyways need to
do this fix to support counters.
> Beam Fn Api metrics piling causes pipeline to stuck after running for a while
> -----------------------------------------------------------------------------
>
> Key: BEAM-8314
> URL: https://issues.apache.org/jira/browse/BEAM-8314
> Project: Beam
> Issue Type: Bug
> Components: runner-dataflow
> Reporter: Yichi Zhang
> Priority: Blocker
> Fix For: 2.16.0
>
> Attachments: E4UaSUhJJKF.png
>
> Time Spent: 40m
> Remaining Estimate: 0h
>
> Seems that in StreamingDataflowWorker we are not able to update the metrics
> fast enough to dataflow service, the piling metrics causes memory usage to
> increase and eventually leads to excessive memory thrashing/GC. And it will
> almost stop the pipeline from processing new items.
>
> !E4UaSUhJJKF.png!
--
This message was sent by Atlassian Jira
(v8.3.4#803005)
