python-memory-leak-analyzer

To diagnose a Python memory leak causing continuous RSS growth, you must correlate process-level memory metrics with Python heap snapshots. This distinguishes whether the root cause is retained objects, native allocator fragmentation, or file-backed mapping expansion.

Your Python service is getting OOM killed due to unbounded memory growth from global container leaks, closures retaining frames, or native C-extension expansion. Analyzing semantic retention chains and native allocation patterns isolates the specific trigger.

Distinguishing native memory leaks from Python object retention requires correlating RSS growth with native allocation patterns and semantic retention chains. This multi-layered diagnosis maps heap snapshots against process-level memory metrics to isolate the leak source.

You can safely analyze memory fragmentation in a production Python service because the diagnostic operates in a read-only mode by default. It accesses the /proc filesystem and standard diagnostic tools without modifying the running process, ensuring production stability.

To troubleshoot Python memory growth using the /proc filesystem, you need access to the target process PID and standard Python diagnostic tools. This allows the analysis to read process-level memory metrics and correlate them with heap snapshots.