Deployment Manual / Version 2506.0
Table Of ContentsThe two collectors make different trade-offs that operators should weigh against their deployment’s traffic profile and operational capabilities:
| Parallel GC (default) | G1GC (opt-in) | |
|---|---|---|
OOM risk | Low — JDK bug does not apply | Elevated on JDK < 25 (see above) |
GC pause duration | Long — seconds on large heaps | Short and bounded by |
Responsiveness under load | Degraded during Full GC | Consistently good |
Raw GC throughput | Higher | Slightly lower due to concurrent GC threads |
Large image transformation | Safe | Risk of heap fragmentation (see tuning note below) |
CoreMedia support | Full (default configuration) | Customer responsibility; not actively supported |
G1GC is worth considering for deployments that run with a well-provisioned heap, have active GC monitoring in place, and can react quickly to an OOM event should one occur. It is especially relevant for:
Content Application Engine (CAE) — live instances serving high-traffic sites
Headless Server — live instances under sustained load
