Java: 'java.lang.OutOfMemoryError: Java heap space' (or 'GC overhead limit exceeded'). How do I tell a too-small heap from a memory leak and fix it?

Cause: the JVM couldn't allocate an object because the heap was full even after garbage collection. Either the maximum heap is too small for the workload (default max heap is a fraction of physical or container memory, often 1/4), or memory is leaking: objects stay referenced and accumulate (unbounded caches, static collections, listeners never removed, loading a huge result set or file into memory at once). 'GC overhead limit exceeded' is the same problem seen earlier: the JVM spends almost all its time collecting garbage while recovering very little. Fix: 1. Look at memory over time first: if heap usage after each GC keeps rising until the crash, it's a leak; if it rises only with load (one big request, a batch job) and is flat otherwise, the heap is too small for that work. 2. Capture evidence: run with -XX:+HeapDumpOnOutOfMemoryError -XX:HeapDumpPath=/some/dir so the JVM writes a heap dump when it fails, or take one with jcmd <pid> GC.heap_dump file.hprof. Open it in a heap analyzer (e.g. Eclipse MAT or VisualVM) and look at the largest retained objects and what keeps them alive. jcmd <pid> GC.class_histogram gives a quick view without a full dump. 3. If the heap is simply too small, raise it with -Xmx (e.g. -Xmx4g), or in containers use -XX:MaxRAMPercentage=75 so the heap follows the container's memory limit. Leave room for non-heap memory (metaspace, threads, direct buffers) or the container gets OOM-killed instead. 4. If it's a leak, fix the code: bound caches (size limits, expiry), remove listeners, stream large data instead of loading it all, close resources, and avoid holding request data in static fields. 5. Build tools (Gradle, Maven) and IDEs run their own JVMs: set org.gradle.jvmargs in gradle.properties or MAVEN_OPTS for build-time OOMs. Verify: with the fix, heap usage after GC levels off under steady load (watch with jstat -gcutil <pid> or your monitoring), and the job completes without OOM. Sources: https://docs.oracle.com/en/java/javase/21/troubleshoot/troubleshooting-memory-leaks.html https://docs.oracle.com/en/java/javase/21/docs/specs/man/java.html