Heap Dump Analyzer

The tool ingests heap dumps from JVM (HPROF, HPROF-like formats) and .NET (CLR dumps), optionally compressed. It builds a memory graph, resolves object references, and aggregates by class. It computes total live memory, per-class counts, and retained sizes. It identifies dominant memory consumers and potential leaks. It supports on-demand scanning and full dump analysis, flags suspicious allocation sites, and maintains stable object identities to enable cross-dump comparisons.
Core outputs include a ranked list of top memory hotspots by retained size, reference paths from GC roots to suspect objects, dominator tree segments, and a concise summary report. It supports filtering by package or namespace, by a time window in which the dump was captured, and baseline comparisons to detect regressions in leaks. Outputs are designed for both human review and automation; human-friendly summaries accompany detailed machine-readable data exports.
The analysis is designed for developers and site reliability engineers who need actionable results rather than raw data. It produces precise metrics and explainable paths showing how a large object remains reachable. Unique value comes from efficient handling of large graphs, streaming parsing, deterministic sorting, and exportable data in JSON or CSV formats in addition to readable summaries. It also highlights GC root types and lifetime estimates for major objects.
Common use cases include debugging memory leaks in long running services, verifying leak fixes across builds, capacity planning for memory-heavy applications, and post-deployment health checks in production environments. Example scenarios include JVM services with a single large ByteBuffer retained by a root, or a web API whose response objects persist via thread-local caches across requests.