perf-profiling

Profile Python applications to identify performance bottlenecks using multiple profiling backends. The toolkit detects CPU hotspots, memory leaks, and inefficient code paths to pinpoint exactly where execution time and resources are consumed.

Tracking memory leaks in a Python script relies on memray to monitor memory allocations. It generates detailed reports that reveal exactly where and why memory accumulates, helping you resolve leaks and reduce overall memory consumption.

Profiling async Python code paths is fully supported using yappi. It measures concurrent execution across coroutines and threads, providing accurate timing data for complex asynchronous workflows without blocking the event loop.

Generate a flame graph for a Python process by running CPU profilers like py-spy or Scalene. They sample stack traces and output visual flame graphs that immediately highlight the most active functions consuming execution time.

Line-by-line timing in Python works effectively with line_profiler. It measures the execution time of individual statements within a function, providing granular insights needed to optimize specific lines of code rather than whole functions.

For CPU profiling, py-spy samples stack traces with minimal overhead, while Scalene analyzes both CPU and memory at the line level. Choose py-spy for low-overhead sampling and Scalene for deeper line-level resource analysis.