Nirmal Chandra

Hi Sneha, apologies for late reply. This was was not accpeting any reply and comments I tried to put earlier. Where exactly you are facing this issue? Did you clone the example repo and run the tests? Did you start the Kafka cluster "docker-compose up" before you ran the test? If you are still facing this, please ping us in the Slack... we will be happy to sort this out soon.

Sorry mate, just seeing this question now and apologies for the late reply.

Number of threads or virtual users whether it is 1000 or 5000 really depends on your laptop or the machine you use and how much resources (e.g. CPU cores and RAM etc) are left free to drive your load tests to run.

e.g. if you have a 16gb with dual-core(2) processor machine and no other high-end processes are running, say your API payload size is 1KB size in average + say you have left 12gb of RAM for your load testing, then you can spin up more threads/virtual-users that 1 core with 2 GB of RAM.

But the same may vary if your API payload size it more or less with any other high processing software is running or not running.

My suggestion would be to try from 90 to 100 threads per secs and keep gradually increasing and see how many v-users it can support and how much further resources you need!

If you need far larger load e.g. 10000 virtual users within few seconds rampup duration, you can fire such load from the similar setup in Cloud e.g. AWS or GCP or Azure where you have very large scaling options. Zerocode runners are very very flexible for that purpose.

Cool stuff! Works as expected.

Hello Kevin, Thanks for the comment. The link works now(just tested). Can you please confirm, you are able to navigate?

Good piece of information.

A typo(ParametrizedType) needs to be corrected for everyone's benifit :

Correct one is below.

ParameterizedType

Full qualified class name : java.lang.reflect.ParameterizedType

Good information, short n precise steps !

Thanks for the reply. That's what I thought and that's how I did(still has issues :-( ). I am using the latest version of the kafka-client. Will it matter if I connect to earlier(very old) version of the Kafka server/brokers? Also I used old version of Kafka client and was able to connect these old kafka servers(but via Zoo keeper details).

Hello mate, in the java client config nowhere `zookeeper` was needed. But for the command line it was needed `--zookeeper ...`. So how did Java client connect to the kafka cluster without it? Or am I missing something? If at all I have to provide the `zookeeper` details how to provide for Java client? Thanks in advance.

thanks for publishing the nice article. It's very clean approach for testing end-points. No more overheads of maintaining POJOs, their builders, serialization/deserialization, boilerplate glue codes and imagining how my test would look like.

Hi,

SOAP calls has been answered here.

DB ResultSet assertion e.g. Here

i.e. basically write a simple Java public method and call it like this.

Regards,

N C (Contributor)

Good approach! Now we have done this for our Mule application (for a Bank). It's convinient and saves time switching environments. `mvn -Dmule.env=` helps now.

Good one, how to capture n process the path n query params