# Kafka performance: RAM

●

●

Published 2022-12-03

This tutorial explains to you the impact of the RAM on the Kafka server performance.

Here are some information to retain about the RAM memory & Kafka cluster performance:

ZooKeeper uses the JVM heap, and 4GB RAM is typically sufficient. Too small of a heap will result in high CPU due to constant garbage collection while too large heap may result in long garbage collection pauses and loss of connectivity within the ZooKeeper cluster.
Kafka brokers use both the JVM heap & OS page cache
the JVM heap : used for replication of partitions between brokers and for log compaction. For small to medium-sized deployments, 4GB heap size is usually sufficient.
OS page cache : the consumers read data from here. OS page cache size depends on the retention period, the number of messages per second and the dimension of the message.
Usually we can see 32Gb or 64Gb RAM on production database machine;
The Kafka JVM heap is set up by using the KAFKA_HEAP_OPTS environment variable. The heap size must be monitored over time and could be increased if we have more partitions in the broker.

Example:

export KAFKA_HEAP_OPTS = "-Xmx4g"

Don't set -Xms (starting heap size) parameter: let it grow naturally.

The OS page cache is not set manually. The remaining RAM is used automatically by OS page cache.
Don't forget to disable RAM Swap - can set to 0 on certain Linux versions/distributions (by default it is 60 on Linux)

sysctl vm.swappiness=1
echo 'vm.swappiness=1' | sudo tee --append /etc/sysctl.conf