Re: [Question] Testing interaction of streaming main input and slowly updating side input

2021-09-13 Thread Siyu Lin
Hi Johannes, L20 in your code will make it run forever so it is better to put a finite number there, like `GenerateSequence.from(0).to(1).withRate(1, Duration.standardSeconds(2))` -siyu > On Sep 10, 2021, at 6:06 AM, Johannes Frey wrote: > > Hi everybody, > > I'm currently having a hard tim

Re: Beam program with Flink runner which can limit number of records processed in specified time when reading from Kafka

2021-12-15 Thread Siyu Lin
Hi Luke, I had the same issue before. Was wondering if it is possible to use a different timestamp as the watermark. Thanks! -siyu On Fri, Nov 19, 2021 at 11:42 AM Luke Cwik wrote: > The issue seems to be that when doing a backfill Kafka is using wall time > to estimate the watermark and then

Java SDK: Is there a system property to print out detailed stack trace in test?

2022-01-24 Thread Siyu Lin
Hi all, Was wondering if we can set up a system property to print out detailed errors when tests throwing exceptions / errors. I remember that I set it up before but forget the exact name. Appreciate any help! -siyu

Re: "Slowly updating global window side inputs" example buggy?

2022-03-28 Thread Siyu Lin
Hi Reza, Just read a blog post[1] by you two years ago and you mentioned > Because this pattern uses a global-window SideInput, matching to elements > being processed will be nondeterministic. Do you mean two workers working on the same windowed main input and use different global windowed si

Histogram metrics in Dataflow/Beam

2022-04-04 Thread Siyu Lin
Hi Beam community, I am wondering if there is histogram metrics available (or alternative recommendations) for showing up quantiles. We have counter metrics already but we would also like to see some quantiles for different values. Thanks a lot! Siyu

Re: Histogram metrics in Dataflow/Beam

2022-04-04 Thread Siyu Lin
epend on the runner. For Cloud Dataflow, the reported > values are MAX, MIN, MEAN, and COUNT, so no support for finer-grained > percentiles: > > https://cloud.google.com/dataflow/docs/guides/using-cloud-monitoring#custom_metrics > >> On Mon, Apr 4, 2022 at 7:45 PM Siyu Lin wr

Breaking change for FileIO WriteDynamic in Beam 2.34?

2022-04-05 Thread Siyu Lin
Hi Beam community, We have a batch pipeline which does not run regularly. Recently we have upgraded to Beam 2.36 and this broke the FileIO WriteDynamic process. We are using Dataflow Runner, and the errors are like this when there are multiple workers: Error message from worker: java.lang.NoClas

Re: IllegalMutationException in PTransform

2022-05-06 Thread Siyu Lin
Hi Reuven, Do you mean we should have coder explicitly defined for all input types and output types in chaining do fns? Do we also need to have comparedTo and equals defined as well? thanks again! Siyu > On May 6, 2022, at 12:23 PM, Reuven Lax wrote: > >  > Could be - I would check the impl

Re: Histogram metrics in Dataflow/Beam

2022-05-06 Thread Siyu Lin
> On May 6, 2022, at 11:57 AM, Pablo Estrada wrote: > > Sorry about the delay! > > Yes, you can add any dependencies to your image - and you can add custom > reporting of metrics that you're tracking directly. That may help? > > On Mon, Apr 4, 2022 at 7

Re: Histogram metrics in Dataflow/Beam

2022-05-06 Thread Siyu Lin
worker - I suppose you can add new > binaries that run next to your Beam worker processes (in the same container, > and you have to start them up in the entrypoint of your Dockerfile). I've > never seen anyone do this, but it sounds quite doable. > > Do any of these opti

Re: Histogram metrics in Dataflow/Beam

2022-05-06 Thread Siyu Lin
sorry for typo “recommissioned way” should be “recommended way” > On May 6, 2022, at 4:31 PM, Siyu Lin wrote: > >  > Hi Pablo, > > Thanks so much for your explanation! > > Also, for prom client in the do fn, do we need to initialize them in the > setup code? My

Re: Histogram metrics in Dataflow/Beam

2022-05-06 Thread Siyu Lin
by many threads at the same time - so you need to make sure that > PrometheusClient is thread safe, or you need to synchronize accesses to it. > (maybe it supports some kind of async call?) > > Let me know if that makes sense? > Best > -P. > >> On Fri, May 6, 2022 at