A core application based on Akka Cluster had ongoing memory issues, requiring to run virtual machines with a lot of memory (64+ GB) as a workaround. I held a performance engineering workshop to show the team how to use the adequate tooling for performance profiling and memory pressure analysis. A key result of the workshop was a tenfold decrease of the JVM memory required to run the project.

Topics & tools: Performance Engineering, Async Profiler, JVM memory management