Hi John,
The default for the C# client is actually 5ms, not that it really makes a
difference. Are you able to share any of your producer code, especially
around the configuration?
Kind regards,
Liam Clarke
On Thu, Dec 3, 2020 at 5:01 PM John Coleman wrote:
> Hi,
>
> We are using Confluent.Ka
Hi,
We are using Confluent.Kafka (dotnet) and according to the docs the linger.ms
is default to 0.5ms which would cover a lot of serializations since each occurs
in < 10 microseconds. A typical transaction batch would be 2 to 3 messages,
occasionally more.
Regards,
John
On 14/11/20 05:24, "G
Hi,
We send small transactions (2 to 3 messages) a few times per second from each
of 5 to 10 producers, depending on the consumer partition size.
Regards,
John
On 14/11/20 03:55, "Guozhang Wang" wrote:
Hello John,
It's a bit hard to reason what's your total producing traffic to the
Kafka Streams uses an incremental computation model, ie, each time a new
message arrives, the current session is updated immediately (note that
the result is a `KTable`). -- Kafka Streams does not accumulate the raw
events and computes the session result when the window closes, but it
refines the s
Eric,
this idea came up multiple time already, but it is much more difficult
to build than it seems, especially for supporting "exactly once". Kafka
transaction only work on a single cluster...
Maybe we will support it at some point, but the "workaround" to either
replicate the input or output to
Thanks for the heads-up.
I agree this should go into 2.7. I'll roll a new RC once you merge the
patch.
-Bill
On Wed, Dec 2, 2020 at 1:28 PM Jason Gustafson wrote:
> Hi Bill,
>
> I found a potential blocker here:
> https://issues.apache.org/jira/browse/KAFKA-10799. A patch should be along
> sh
Hi Bill,
I found a potential blocker here:
https://issues.apache.org/jira/browse/KAFKA-10799. A patch should be along
shortly.
Best,
Jason
On Mon, Nov 30, 2020 at 7:27 AM Bill Bejeck wrote:
> Thanks for the vote, Gwen.
>
> Here's an update for Jenkins build
>
> * Successful Jenkins builds fo
Subscribe to Kafka dev mailing list
