joerick/pyinstrument
Pyinstrument
pyinstrument is a statistical sampling profiler for Python that records the call stack at regular intervals to identify performance bottlenecks with low overhead. It tracks wall-clock time, including I/O and external service calls, and provides specialized profiling for asynchronous programs by attributing time spent awaiting tasks to the calling function.
The project converts captured execution data into interactive HTML reports, JSON, and flamecharts. It includes a call stack visualizer to simplify the analysis of execution paths and supports the profiling of individual cells within interactive notebooks.
The tool covers a broad range of analysis capabilities, including C-function execution tracking, session management for saving and reloading data, and the ability to filter out irrelevant library frames using glob patterns. It can be used via a programmatic interface with decorators and context managers, or as a command-line tool to analyze scripts and web requests.
Features
- Python Profilers - Identifies slow functions and bottlenecks in Python code by sampling the call stack and measuring wall-clock time.
- Execution Stack Samplers - Identifies performance bottlenecks by sampling the call stack at regular intervals with low overhead.
- Async Timing Attribution - Tracks time spent awaiting asynchronous tasks to accurately profile code execution across different contexts.
- Runtime Call-Stack Capturers - Captures the runtime call stack to identify the source of execution delays.
- Asynchronous Profilers - Tracks execution context in asynchronous programs to attribute time spent awaiting tasks.
- Asynchronous Attribution - Attributes time spent awaiting asynchronous tasks to the calling function to ensure accurate profiles in concurrent programs.
- Asynchronous Programming Patterns - Tracks time spent awaiting tasks in async Python code to understand where an event loop is blocked.
- Wall-Clock Time Measurement - Tracks total elapsed time including I/O operations and external service calls.
- Function Execution Timing - Records the execution time of functions and modules to identify performance bottlenecks.
- Call Graph Visualizers - Converts captured execution data into interactive HTML reports, JSON, and flamecharts for performance analysis.
- Statistical Profilers - Uses a sampling profiler that records the Python call stack at regular intervals to identify performance bottlenecks with low overhead.
- Wall-Clock Time Profilers - Tracks total elapsed time including I/O and external service calls rather than just CPU cycles.
- Application Debugging - Finds the root cause of execution delays in scripts or modules to improve overall program speed.
- Call Tree Aggregators - Groups repeated function calls into a hierarchical tree structure to simplify the analysis of complex execution paths.
- CLI Profiling Tools - Enables executing scripts from the terminal to generate immediate performance reports.
- Code Instrumentation Utilities - Provides a programmatic interface to instrument functions and methods for performance monitoring.
- External Script Analyzers - Analyzes scripts, modules, or code strings from the terminal without requiring changes to original source files.
- Cell-Level Profiling - Allows profiling specific code cells within interactive notebooks using specialized commands.
- Notebook Profilers - Provides a performance measurement tool for analyzing execution time of individual cells within interactive Python notebooks.
- Notebook Performance Optimization - Measures the execution time of specific cells within interactive notebooks to optimize data science workflows.
- Script Performance Analysis - Analyzes execution time for scripts run from the command line to summarize time distribution.
- Internal Frame Filters - Removes noise from performance reports by excluding specific library or system frames using glob pattern matching.
- Web Request Profiling - Analyzes the call stack of HTTP requests to generate performance reports for web-based traffic.
- Code Block Timing - Measures elapsed time for specific segments of code using context managers or decorators.
- Performance Metric Exports - Exports captured performance data into HTML, JSON, and trace files for external visualization.
- Profiling Visualization Filters - Hides irrelevant library frames and modules using glob patterns to focus on application code.
- Web Request Profilers - Analyzes the call stack of HTTP requests to find slow operations in web applications and APIs.
- Performance Profiling - Call stack profiler for Python code.
- System Monitoring And Profiling - Call stack profiler that identifies performance bottlenecks.
