I took a look at the code and nothing obvious stood out to me in the code
as this is a ParDo with OnWindowExpiration. Just to make sure, the rate
limit is per key and would only be a global rate limit if there was a
single key.

Are the workers trying to start?
* If no, then you would need to open a support case and share some job ids
so that someone could debug internal service logs.
* If yes, then did the workers start successfully?
** If no, logs should have some details as to why the worker couldn't start.
** If yes, are the workers getting work items?
*** If no, then you would need to open a support case and share some
job ids so that someone could debug internal service logs.
*** If yes then the logs should have some details as to why the work items
are failing.


On Fri, Aug 5, 2022 at 7:36 AM Evan Galpin <[email protected]> wrote:

> Hi all,
>
> I'm trying to create a RateLimit[1] transform that's based fairly heavily
> on GroupIntoBatches[2]. I've been able to run unit tests using TestPipeline
> to verify desired behaviour and have also run successfully using
> DirectRunner.  However, when I submit the same job to Dataflow it
> completely fails to start and only gives the error message "Workflow
> Failed." The job builds/uploads/submits without error, but never starts and
> gives no detail as to why.
>
> Is there anything I can do to gain more insight about what is going
> wrong?  I've included a gist of the RateLimit[1] code in case there is
> anything obvious wrong there.
>
> Thanks in advance,
> Evan
>
> [1] https://gist.github.com/egalpin/162a04b896dc7be1d0899acf17e676b3
> [2]
> https://github.com/apache/beam/blob/c8d92b03b6b2029978dbc2bf824240232c5e61ac/sdks/java/core/src/main/java/org/apache/beam/sdk/transforms/GroupIntoBatches.java
>

Reply via email to