# Throughput &metrics

**URL:** <https://forum.confluent.io/t/throughput-metrics/36534>\
**Category:** Kafka Streams\
**Created:** [24 January 2025 12:52 UTC](https://forum.confluent.io/t/throughput-metrics/36534 "2025-01-24T12:52:53Z")\
**Posts on this page:** 1\
**Page:** 1

<div class="post-metadata">

**Author:** ![MikeMark](https://avatars.discourse-cdn.com/v4/letter/m/da6949/32.png) [@MikeMark](https://forum.confluent.io/u/MikeMark)\
**Post date:** [24 January 2025 12:52 UTC](https://forum.confluent.io/t/throughput-metrics/36534/1 "2025-01-24T12:52:53Z")

</div>

Hello everyone . I am working on a machine learning engine using kafka and kafka streams( vanilla kafka project) .  
I am a little bit confused about throughput.Right now i send arround 700000 records and i got ( 30000 records/second ) into my kafka streams class with 4 streams threads. If i send more data like 3x initial data my throughput is even better ( 50000 records /second)  
How to use metrics in general? How to simulate an optimal scenario, How i would be sure that i am doing well when eveything changes everytime?