The production-scale datacenter profiler (C/C , Go, Rust, Python, Java, NodeJS, .NET, PHP, Ruby, Perl, …) - elastic/otel-profiling-agent

On using JFR Analytics to locate a thread leak on a java application.

In which scratching the surface of hashCode() leads to a speleology trip through the JVM source reaching object layout, biased locking, and surprising performance implications of relying on the default hashCode().

Skija a collection of hand-crafted bindings for the JVM to the Skia library (the one that powers Chrome, Android, Flutter and Xamarin).

When tuning a prod Scala deployment a while back, I encountered a nasty “pregnant pause” every once in a while similar to what you have experienced. So, below are what I used in annotated form and adapted for the deployment you describe (the max heap setting). Some of these may be completely obvious to you (or others reading), yet are included for completeness.

UseG1GC

Surprisingly common scenarios:

• requiring 1G memory to support a few hundred users
• saving 500K session state per user
• requiring 2M for a text index per simple document
• creating 100K temporary objects per user hit

This covers the basics of memory allocation and garbage collection in Java, the different garbage collectors available in HotSpot and how they can be tuned, and finally some anecdotes from Attila’s experiences at Twitter.