All 15 Debugging & Profiling questions
More in Debugging & Profiling
How do you use cProfile to identify which functions are consuming the most time in a larger program? How do assert statements work for debugging, and why should they NOT be relied upon for production input validation? How do you use Python's built-in 'warnings' module to surface deprecated or risky code paths during debugging? How would you use the 'faulthandler' module to diagnose a program that crashes with a segmentation fault, typically from a C extension? How would you use the 'logging' module's exception logging (logger.exception()) INSIDE an except block to automatically capture the full traceback? How do you use pdb's post-mortem debugging mode to inspect a program's state exactly at the point where an unhandled exception occurred? How would you profile memory allocations specifically (not just CPU time) to find which objects are consuming the most memory in a running process? How do you use conditional breakpoints in pdb to pause execution only when a specific condition is met, avoiding manually stepping through many unrelated iterations? How would you use line_profiler (a third-party tool) to get LINE-BY-LINE timing information within a specific function, more granular than cProfile's function-level data? How would you use py-spy (a sampling profiler) to profile a running Python process in PRODUCTION without restarting it or adding any code changes?
Advertisement
Python › Debugging & Profiling

How do you use memory_profiler (or tracemalloc from the standard library) to identify a memory leak's source in a Python application?

Advanced

Answer

tracemalloc.start() begins tracking memory allocations by source location; taking two snapshots (tracemalloc.take_snapshot()) before and after a suspected leaking operation, then comparing them with snapshot2.compare_to(snapshot1, 'lineno'), reveals exactly which lines of code are responsible for growing memory usage.
import tracemalloc

tracemalloc.start()
snapshot1 = tracemalloc.take_snapshot()
run_suspected_leaking_operation()
snapshot2 = tracemalloc.take_snapshot()

top_stats = snapshot2.compare_to(snapshot1, "lineno")
for stat in top_stats[:5]:
    print(stat)
Real-world example Pinpointing the exact line of code responsible for a service's steadily growing memory usage over time.

Common follow-ups: How does tracemalloc's overhead compare to running a full external memory profiler, in terms of production suitability?

Memory Management & Garbage Collection
All 15 Debugging & Profiling questions