What did you do when your system was slow due to GC?

When experiencing performance issues due to garbage collection (GC) in Java, I usually take the following steps:

1. Memory and GC profiling — I use tools like VisualVM, JProfiler, or Java Flight Recorder to understand what objects are being created and how often GC occurs.

2. Code optimization — I reduce the number of temporary objects, avoid frequent creation of large arrays or strings, and use object pools if justified.

3. JVM parameter tuning — I select an appropriate GC (e.g., G1, ZGC, Shenandoah), configure heap sizes, young and old generation sizes, pause times, and collection frequency.

4. Using weak references and caches — to avoid holding onto unnecessary objects.

5. Algorithm refactoring — sometimes helps reduce memory load.

Example of JVM tuning for G1 GC with pause time limit:

java -XX:+UseG1GC -XX:MaxGCPauseMillis=200 -Xmx2g -Xms2g -jar app.jar

In summary, it’s a comprehensive approach: first analysis, then code and JVM configuration optimization.