Tracking Memory Access Patterns in Python Using sys.settrace and Custom Hooks

Python doesn’t expose raw memory the way lower-level languages do, but you can still track how and when your variables change. In this tutorial, we’ll build a lightweight memory tracing tool using Python’s built-in sys.settrace function, allowing us to monitor variable assignments and get insights into runtime behavior — no C extensions, no third-party tools.

Setup

This method works in pure Python 3. No dependencies needed. You’ll just need a working Python environment and a target function you want to introspect. It's particularly useful for debugging, reverse engineering, or building your own lightweight instrumentation tools.

Full Working Code

import sys

def trace_vars(frame, event, arg):

    if event != "line":

        return trace_vars

code = frame.f_code
lineno = frame.f_lineno
locals_now = frame.f_locals.copy()
global last_locals

if code.co_name not in last_locals:
    last_locals[code.co_name] = locals_now
    return trace_vars

old_locals = last_locals[code.co_name]

for var, new_val in locals_now.items():
    if var not in old_locals:
        print(f"[{code.co_name}:{lineno}] NEW {var} = {new_val}")
    elif old_locals[var] != new_val:
        print(f"[{code.co_name}:{lineno}] MODIFIED {var}: {old_locals[var]} → {new_val}")

for var in old_locals:
    if var not in locals_now:
        print(f"[{code.co_name}:{lineno}] DELETED {var}")

last_locals[code.co_name] = locals_now
return trace_vars



    

    




def monitor(func):

    def wrapper(*args, **kwargs):

        global last_locals

        last_locals = {}

        sys.settrace(trace_vars)

        try:

            return func(*args, **kwargs)

        finally:

            sys.settrace(None)

    return wrapper

  
  
  Example usage


@monitor

def run_example():

    a = 10

    b = a + 5

    b = b * 2

    del a

    return b

run_example()

Explanation

By using sys.settrace, we register a line-by-line callback that can introspect the current local variables at each step. We store a snapshot of the locals for each function, then compare it on the next invocation to detect changes — additions, updates, deletions. It’s a powerful (and often overlooked) way to understand the control and data flow of any Python function at runtime.

Pros & Cons

✅ Pros

Works with unmodified Python
Fully userland — no C or unsafe operations
Easy to wrap around any function
Valuable for introspection, testing, and education

⚠️ Cons

Significant performance overhead
Doesn’t work well with async or multithreaded code
Can’t access native extensions or memory directly
Not suitable for production monitoring

Summary

This approach gives you a window into the lifecycle of variables in Python, letting you trace changes without any third-party packages or tooling. While it's not a true memory inspector, it's a powerful pattern for anyone debugging tricky code or trying to understand execution in fine-grained detail.

If this was useful, you can Buy Me a Coffee ☕

HexShift @hexshift