navidnabavi/profile_function

An easy profiling tool which collects elapsed time for functions and code blocks.

Project Description

A simple profiling tool which collects elapsed time for functions and code blocks.

Usage

Using statsd:

from profile_function import ProfileFunction, StatsdBackend
import statsd

statsd_client = statsd.StatsClient('localhost', 8125)
pf = ProfileFunction(StatsdBackend(statsd_client))

@pf.profile_function(group="rpc")
def f(x, y):
    s = 0
    with pf.profile_block("for-loop"):
        for i in range(x):
           s += i * x + y 
    return s

Using Prometheus:

from profile_function import ProfileFunction, PrometheusBackend
from prometheus_client import start_http_server

start_http_server(9090)

pf = ProfileFunction(PrometheusBackend())

@pf.profile_function(group="rpc")
def f(x, y):
    s = 0
    with pf.profile_block("for_loop"):
        for i in range(x):
           s += i * x + y
    return s

Using logger:

from profile_function import ProfileFunction, LoggerBackend
import logging

logger = logging.getLogger(__name__)
pf = ProfileFunction(LoggerBackend(logger, log_level=logging.DEBUG))

@pf.profile_function(group="rpc")
def f(x, y):
    s = 0
    with pf.profile_block("for-loop"):
        for i in range(x):
           s += i * x + y 
    return s

Install

You can install profile_function using pip:

$ pip install profile-function

If you are using statsd you need to install it first. This project does not mentioned statsd as it own dependency. This command may be useful then:

$ pip install statsd

Development:

You can write your own collector if you need by implementing CollectorBackend.