Confluent Cloud Q1 Launch: Build a Secure Shared Services Data Streaming Platform | Learn more

The secret life of Kafka JVM Heap

In this session I will talk about our experience with the Java Garbage collector and Kafka in production. The talk will have a brief introduction to G1GC, some garbage collector metrics to keep an eye for, how to track GC efficiency, linux page cache, possible parameters to tweak, typical throughput and latency that we have seen and if time permits some experiments with a future garbage collector (ZGC).

Presenter

Adithya Chandra

Adithya Chandra is a software engineer on the Kafka Performance team, where he is focused on making Apache Kafka more scalable and performant on Confluent Cloud. Before Confluent he worked at Amazon Web Services (AWS) on search services and Amazon Aurora on performance, load-balancing, cluster management and elasticity of search and storage clusters.