The story of a Java 17 native memory leak

Voting no longer possible
Voting enabled when talk has started

Shortly after the release of Java 17 the platform team in my organisation prepared a new Docker base image that allowed product teams to upgrade and create new services that target the latest LTS. Within a few days of some services being upgraded we spotted some interesting trends in respect of the memory usage for them. Simply switching the JDK from 11 to 17 had resulted in a slow native memory leak!

This ignite talk charts the story of how a Java 17 native memory leak played out in public on Twitter demonstrating the power of the Java community. We'll follow the journey through discovery, investigation, confusion, enlightenment, resolution and finally verification of the fix.

Nick Ebbitt

Morgan Stanley

Nick is a software engineer working with various technologies in multiple industries for over 20 years. He is a developer with a strong interest in operations and a passion for building platforms, enabling product teams and developers to become more effective at delivering change. His passion for Java and the community have led him to become leader of his local JUG, the Manchester Java Community.