Performance tuning Java

Last updated 11th May 2021

Objective

There are a number of settings that can be adjusted for each application to optimize its performance on Web PaaS.

Memory limits

The JVM generally requires specifying a maximum memory size it is allowed to use, using the Xmx parameter. That should be set based on the available memory on the application container, which will vary with its size.

To extract the container-scaled value on the command line, use $(jq .info.limits.memory /run/config.json).

You should also set the ExitOnOutOfMemoryError. When you enable this option, the JVM exits on the first occurrence of an out-of-memory error. Web PaaS will restart the application automatically.

These are the recommended parameters for running a Java application. Thus, the command to use to start a Java application is:

java -jar -Xmx$(jq .info.limits.memory /run/config.json)m -XX:+ExitOnOutOfMemoryError //The rest of the arguments and the jar file.

Garbage collection

When migrating the application to a cloud environment, it is often essential to analyze the Garbage Collector's log and behavior. For this, there are two options:

  • Placing the log into the Web PaaS /var/log/app.log file (which captures STDOUT).
  • Creating a log file specifically for the GC.

To use the STDOUT log, you can add the parameter -XX: + PrintGCDetails, E.g.:

java -jar -Xmx$(jq .info.limits.memory /run/config.json)m -XX:+ExitOnOutOfMemoryError -XX:+PrintGCDetails //The rest of the arguments and the jar file.

Java supports a number of different garbage collection strategies. Which one is optimal for your application will vary depending on your available memory, Java version, and application profile. Determining which is best for your application is out of scope, but the main options and how to enable them are:

Name Command Flag Description
Serial Garbage Collector -XX:+UseSerialGC This is the simplest GC implementation, as it basically works with a single thread.
Parallel Garbage Collector -XX:+UseParallelGC Unlike Serial Garbage Collector, this uses multiple threads for managing heap space. But it also freezes other application threads while performing GC.
CMS Garbage Collector -XX:+USeParNewGC The Concurrent Mark Sweep (CMS) implementation uses multiple garbage collector threads for garbage collection. It's for applications that prefer shorter garbage collection pauses, and that can afford to share processor resources with the garbage collector while the application is running.
G1 Garbage Collector -XX:+UseG1GC Garbage First, G1, is for applications running on multiprocessor machines with large memory space.

The default strategy on Java 9 and later is G1. The GC strategy to use can be set in the start line with:

Serial

java -jar -Xmx$(jq .info.limits.memory /run/config.json)m -XX:+ExitOnOutOfMemoryError -XX:+PrintGCDetails -XX:+UseSerialGC //The rest of the arguments and the jar file.

Parallel Garbage Collector

java -jar -Xmx$(jq .info.limits.memory /run/config.json)m -XX:+ExitOnOutOfMemoryError -XX:+PrintGCDetails -XX:+UseParallelGC //The rest of the arguments and the jar file.

CMS Garbage Collector

java -jar -Xmx$(jq .info.limits.memory /run/config.json)m -XX:+ExitOnOutOfMemoryError -XX:+PrintGCDetails -XX:+USeParNewGC //The rest of the arguments and the jar file.

G1

java -jar -Xmx$(jq .info.limits.memory /run/config.json)m -XX:+ExitOnOutOfMemoryError -XX:+PrintGCDetails -XX:+UseG1GC //The rest of the arguments and the jar file.

Java 8 Optimization

Ideally, all applications should run the latest LTS release of the JVM at least. That is currently Java 11. Java 11 has a number of performance improvements, particularly on container-based environments such as Web PaaS.

However, in many cases, this is not possible. If you are still running on Java 8 there are two additional considerations.

The default garbage collector for Java 8 is Parallel GC. In most cases G1 will offer better performance. We recommend enabling it, as above.

Furthermore, there is the UseStringDeduplication flag which works to eliminate duplicate Strings within the GC process. That flag can save between 13% to 30% of memory, depending on application. However, this can impact on the pause time of your app.

java -jar -Xmx$(jq .info.limits.memory /run/config.json)m -XX:+UseG1GC -XX:+UseStringDeduplication -XX:+ExitOnOutOfMemoryError -XX:+PrintGCDetails

References


Did you find this guide useful?

Please feel free to give any suggestions in order to improve this documentation.

Whether your feedback is about images, content, or structure, please share it, so that we can improve it together.

Your support requests will not be processed via this form. To do this, please use the "Create a ticket" form.

Thank you. Your feedback has been received.


These guides might also interest you...

OVHcloud Community

Access your community space. Ask questions, search for information, post content, and interact with other OVHcloud Community members.

Discuss with the OVHcloud community

In accordance with the 2006/112/CE Directive, modified on 01/01/2015, prices exclude VAT. VAT may vary according to the customer's country of residence.