I want to do something crazy. Someone talk me off the ledge:
record Bundle {
string key;
array msgs;
}
Producers individually serialize a bunch of messages that share a key, then
serialize a bundle and post to a topic.
A generic Flattener service is configured by startup parameters to liste
Hello Nan,
Note that Streams may need some time to rebalance and assign tasks even if
you only starts with one instance.
I'd suggest you register your state listener in Kafka Streams via
KafkaStreams#setStateListener, and your customized StateListener should
record when the state transits from RE
Hi Guozhang,
Please find below. I have tried with the latest 2.0.0 libraries and no
improvement observed.
Kafka version - 1.0.1
Total Memory allocated - 24 GB
Max Stream Cache - 8GB
---
Processor class code:
private KeyValueStore hourlyStore = null; // Loca
right, so my kafka cluster is already up and running for a while, and I can
see from the log all broker instance already change from rebalance to
running.
I did a another test.
from producer, right before the message get send to the broker, I put a
timestamp in the message. and from the consumer s
Okay, so you're measuring end-to-end time from producer -> broker ->
streams' consumer client, there are multiple phases that can contribute to
the 100ms latency, and I cannot tell if stream's consumer phase is the
major contributor. For example, if the topic was not created before, then
when the b
How many brokers are there in your cluster? This error usually comes when
one of the brokers who is leader for a partition dies and you are trying to
access it.
On Fri, Aug 17, 2018 at 9:23 PM Harish K wrote:
> Hi,
>I have installed Kafka and created topic but while data ingestion i get
> so
Firstly, record size of 150mb is too big. I am quite sure your timeout
exceptions are due to such a large record. There is a setting in producer
and broker config which allows you to specify max message size in bytes.
But still records each of size 150mb might lead to problems with increasing
volum
did more test and and make the test case simple.
all the setup now is a single physical machine. running 3 docker instance.
a1, a2, a3
kafka + zookeeper running on all of those docker containers.
producer running on a1, send a single key, update speed 2000 message/s,
each message is 10K size.
3 c
I did several test. one is with 10 brokers (remote server),
one with 3 brokers. (local docker)
both exhibit the same behavior, I was thinking the same but from at least
the kafka log, I don't see a rebalance happening. and I am sure my cpu is
only used about half. and all broker still running.
maybe I should highlight, I only publish 1 key. so only one broker is going
to handle it. and only 1 stream instance handle it too. what's the typical
throughput/latency I should expect in this case? assuming the processing
logic is very very simple, just get data(integer) and sum. I am more
expect
10 matches
Mail list logo