Memory Leak Investigation overview

Systematically diagnoses and resolves memory leaks in production or staging services -- capturing heap snapshots over time, analyzing object retention graphs, identifying leaked allocations by class and call site, correlating memory growth with traffic patterns or specific endpoints, profiling native memory via tools like jemalloc or valgrind, differentiating true leaks from excessive caching, producing a root-cause report with fix recommendations, and verifying the fix stabilizes RSS under sustained load. Excludes preventive memory budgeting.