openclaw-test-heap-leaks

Compare heap snapshots from the same worker PID across multiple test lanes to identify whether memory growth stems from transformed-module retention or runtime leaks.

Your pnpm test run may go OOM due to shared-worker memory growth or retained objects; collecting repeated .heapsnapshot files helps triage whether the growth is from transformed-module retention or runtime leaks.

Use the included heap snapshot delta tooling to parse .heapsnapshot files, summarize sizes by type and name, and report top deltas with before/after references to isolate retention hotspots.

Yes, heap snapshot analysis applies to both CI test runs and local development with pnpm test across multiple workers, collecting .heapsnapshot files to establish reproducible memory growth patterns.

The best way to triage OOM causes in OpenClaw tests is analyzing heap snapshots from the same worker PID to distinguish between transformed-module retention and runtime leaks.

Isolate memory hotspots by comparing before and after heap snapshots from the same PID, then patch cleanup logic or test scaffolding based on the reported top size deltas to reduce footprint.