> >>>>>>> >>> expected to
> >>>>>>> >>> have as many
> >>>>>>> &g
Sorry for starting a new thread. I couldn't add a response to the existing
thread at [1] for some reason. I have a very naive question related to
finding a temporary fix to the memory issue raised in [1]. I know Jan
suggested to use 2 successive fixed overlapping windows with offset as a
temporary
it takes for memory usage to reach the limit. Could you also say how
> long it currently takes for memory to reach the limit?
>
> On Thu, Aug 6, 2020 at 4:23 PM David Gogokhiya wrote:
>
> > Hi,
> >
> > We recently started using Apache Beam version 2.20.0 running on
r's
memory usage is an expected behavior or not?
Sincerely,
David
On Thu, Aug 6, 2020 at 3:19 PM David Gogokhiya wrote:
> Hi,
>
> We recently started using Apache Beam version 2.20.0 running on Flink
> version 1.9 deployed on kubernetes to process unbounded streams of data.
> Howeve
Hi,
We recently started using Apache Beam version 2.20.0 running on Flink
version 1.9 deployed on kubernetes to process unbounded streams of data.
However, we noticed that the memory consumed by stateful Beam is steadily
increasing over time with no drops no matter what the current bandwidth is.
W
