Created
January 13, 2022 08:24
-
-
Save mostafa-asg/8817c6eb45d7eb029535dfa34a7003ed to your computer and use it in GitHub Desktop.
How many partitions to use for a topic?
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
The formula for determining the number of partitions per Kafka topic has been pretty well explored over time. | |
When creating a new topic in your Kafka cluster, you should first think about your desired throughput (t) in MB/sec. | |
Next, consider the producer throughput that you can achieve on a single partition (p)—this is affected by producer | |
configurations but generally sits at roughly 10s of MB/sec. Finally, you need to determine the consumer throughput (c) | |
you will have—this is application-dependent so you’ll have to measure it yourself. You should anticipate having at least | |
max(t/p, t/c) partitions for that topic. So if you have a throughput requirement of 250 MB/sec, with a producer and | |
consumer throughput of 50 MB/sec and 25 MB/sec respectively. | |
Then you should use at least max(250/50, 250/25) = max(5, 10) = 10 partitions for that topic. | |
https://www.confluent.io/blog/5-common-pitfalls-when-using-apache-kafka/ |
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment