Sean Glover
Sean Glover
https://travis-ci.org/github/akka/alpakka-kafka/jobs/718740561#L810
Possibly fixed by #1109.
This is actually a timeout of a `waitUntilConsumerSummary` call (10 second timeout). https://travis-ci.com/github/akka/alpakka-kafka/jobs/480442181#L3229
I did some profiling of the Alpakka Kafka Producer benchmark test. Heap memory usage was consistent, but the process only ran for about 5 minutes. ``` benchmarks/it:testOnly *.AlpakkaKafkaPlainProducer -- -z...
Sorry, I just noticed that OP did provide a reproducer. I'll take another look with that impl.
Reproducing benchmarks results is always a challenge. The numbers from the post were not suggesting a baseline performance metric. The point was to visualize the before and after metrics relative...
You haven't provided enough information for me to give you any advice. How large is your local cluster, how many partitions are you using, what is the topic replication factor,...
I used this [benchmark test](https://github.com/akka/alpakka-kafka/blob/master/benchmarks/src/it/scala/akka/kafka/benchmarks/Benchmarks.scala#L88..L109) to produce the results for the blog post. It will output a csv with rows of values representing metrics from brokers and the Kafka consumer...
Assuming the `CommitterSettings` are the defaults and your `config.batchWindowSec` is less than an hour then the only things I can think of is that a stage downstream of `Committer.flow` is...
> my kafka cluster has 6 brokers, and it's idle, only runnings this app, nothing else, why would the Commiter.flow backpressure ? also why for hours? I mean a stage...