Deployment Manual / Version 2506.0
Table Of ContentsCorrect heap sizing is as important as the GC algorithm choice. Two common mistakes lead to OOM incidents in opposite directions:
Heap too small — ParallelGC thrashes and triggers
GC overhead limit exceededunder load.Heap too large — the JVM heap plus off-heap memory (native buffers, Metaspace, thread stacks) exceeds the container memory limit, causing the container to be killed by the OS or Kubernetes.
As a guideline:
Allocate 70–80 % of the container memory limit as Java heap via
JAVA_HEAPor-XX:MaxRAMPercentage. The remaining 20–30 % covers off-heap memory.A heap of 4 GB is a tested baseline for a delivery CAE or Headless Server without heavy customisation. Scale up for higher traffic or more complex business logic.
