Skip to content
This repository has been archived by the owner on Mar 22, 2022. It is now read-only.

Statistics Information about gatling-kafka load generation? #4

Open
daluu opened this issue Oct 1, 2015 · 0 comments
Open

Statistics Information about gatling-kafka load generation? #4

daluu opened this issue Oct 1, 2015 · 0 comments

Comments

@daluu
Copy link

daluu commented Oct 1, 2015

More of a discussion topic and request here.

I don't think this project currently mentions any statistics or information regardling kafka load generation with gatling so far.

Kafka is a relatively new technology, and as far as I recall any mentions about load/performance testing come to articles like these:

https://engineering.linkedin.com/kafka/benchmarking-apache-kafka-2-million-writes-second-three-cheap-machines

https://grey-boundary.io/load-testing-apache-kafka-on-aws/

which seem to be more of benchmarking or generic stressing of a kafka environment without much regard for custom messages and other requirements specific to an organization. Your tool helps satisfy the need for the customizations needed for load testing a kafka-based environment.

It would be nice to know of some stats regarding usage of this tool (as a point of reference) such as what amount of load has been generated by this plugin like:

  • number of kafka messages sent per second
  • kafka message sizes (in bytes) used in the load
  • number of producers/users generating the load
  • load generation schemes (ramp up/down patterns)
  • how much load generated by gatling-kafka on some given setup (single box hardware, distributed test setup) before gatling/gatling-kafka crapped out from lack of resources OR the generated load was good enough for you to not need to scale further
  • how much load generated against the system under test by gatling-kafka

Other users of gatling-kafka (if there are any so far) could also chime in on this topic. I would provide my input once I've used gatling-kafka for some load testing. For now, I'm focusing on kafkameter (for JMeter) as a fallback first since I'm more familiar with that than gatling. Will move on to try gatling-kafka afterwards if I need more "scale" for load generation.

Sign up for free to subscribe to this conversation on GitHub. Already have an account? Sign in.
Labels
None yet
Projects
None yet
Development

No branches or pull requests

1 participant