Optimizing Memory Management and Garbage Collection Overhead in Cloud-Native Java Applications

Abstract

The transition of Java applications from monolithic, bare-metal enterprise deployments to cloud-native, containerized environments (e.g., Kubernetes) has introduced severe memory management challenges. Traditional Java Virtual Machine (JVM) heuristics—designed for high-throughput, multi-gigabyte monolithic runtimes—often conflict with container memory limits, leading to out-of-memory (OOM) kills, CPU throttling, and excessive garbage collection (GC) latency spikes. This paper investigates strategies for optimizing memory management and minimizing GC overhead in cloud-native Java applications. We analyze modern garbage collection algorithms (G1GC, Generational ZGC, and Shenandoah), evaluate native compilation via GraalVM, explore the memory impact of Virtual Threads (Project Loom), and formulate container-aware tuning strategies for modern cloud environments.