Blackfire Continuous Profiler for Python

Blackfire Continuous Profiler continuously collects and uploads profiling data to the Blackfire servers. Once enabled, the profiler collects the relevant profiling information in configurable intervals and periodically uploads it to the Blackfire Agent. Blackfire Agent then forwards this information to the backend.

How to use

Prerequisites

• Python >= 3.7.0
• Blackfire Agent >= 2.13.0

Installation

pip install blackfire_conprof

Example

An example using the whole API interface:

1. Install dependencies

pip install blackfire_conprof

2. Create example.py with the following code

from blackfire_conprof.profiler import Profiler

def foo():
    import time
    time.sleep(1.0)

profiler = Profiler(application_name="my-python-app", agent_socket="tcp://127.0.0.1:8307", labels={'my-extra-label': 'data'})
profiler.start()
foo()
profiler.stop()

3. Run Blackfire Agent (version 2.13.0 and up)

BLACKFIRE_SOCKET="tcp://127.0.0.1:8307" blackfire agent --log-level=4

4. Run the example application. (python example.py)
5. Profiler will send data to the Agent, and Agent will forward it to the Blackfire backend. Data then can be visualized at https://blackfire.io

Enabling the profiler

There are two ways to enable the profiler:

• via Code,
• via Command Line.

Code

from blackfire_conprof.profiler import Profiler

profiler = Profiler()
profiler.start()

If you are working with gevent or any other library that uses pseudo-threads, and you also need to call

import blackfire_conprof.auto

as early as possible in your application. Example: for gevent applications to work correctly, this needs to be imported before gevent.monkey.patch_all().

Command line

Run your Python application as following:

blackfire-conprof python app.py

This will automatically enable the profiler just before your application runs.

Contributing

Use make help to display an overview of useful commands for your dev environment.