Python Module Introduction-multi-mechanize Universal Performance testing Tool

Brief introduction

Multi-mechanize is an open-source performance and load testing framework that concurrently runs multiple Python scripts to generate load on a Web site or service (a composite transaction). The test output report is saved as HTML or JMeter-compatible XML. Multi-mechanize is most commonly used for Web performance and Scalability (scalability) testing, and also for any API that Python can access. Especially suitable for background performance testing. With a little bit of programming, this tool will be far stronger than commercial performance testing tools.

Key Features:

• Supports various HTTP methods

• Advanced hyperlinks and HTML form support

• Support SSL

• Automatic processing of Cookies

• HTTP headers can be set

• Automatically handle redirects

• Support Agent

• Support HTTP Authentication

Installation

Use the standard Python installation method. Note that you need to install Matplotlib to support the drawing, so install yum-y install python27-matplotlib under CENTOS6. Multi-mechanize uses the standard Python installation method pip install Multi-mechanize or Easy_install multi-mechanize here are Linux (CentOS) as an example.

• CentOS 5 Installation Note:

• Note Installation dependencies: Python26 python26-devel python26-numpy* Python26-tornado

• To Epel support: http://dl.fedoraproject.org/pub/epel/5/x86_64/epel-release-5-4.noarch.rpm

Quick Start

Create a project

# Multimech-newproject My_project

Execute project

# Multimech-run My_project user_groups:2 threads:6[================100%==================] 30s/30s transactions:1 timers:119 errors:0waiting for all requests to finish...analyzing Results...transactions:125errors:0test Start:2 013-09-13 11:47:47test finish:2013-09-13 11:48:16created:./my_project/results/results_2013.09.13_11.47.46/ Results.htmldone.

Test results See:

Directory structure

Each test project contains the following content:

• The config.cfg configuration file. Used to set test options.

• test_scripts/the directory of virtual user scripts. Add your test script here.

• results/: The results Store directory. For each test it is claimed that a timestamp directory contains the results of the report.

multimech-newproject, a random number script is generated by default. The script v_user.py as follows:

Import Randomimport Timeclass Transaction (object): Def __init__ (self): Pass def Run (self): R = Random. Uniform (1, 2) time.sleep (r) self.custom_timers[' Example_timer '] = rif __name__ = = ' __main__ ': trans = Tra Nsaction () Trans.run () print trans.custom_timers

• Run_time:duration of test (seconds) execution time

• Rampup:duration of user rampup (seconds) How many seconds Inside Out request

• Results_ts_interval:time series interval for results analysis (seconds) results analyze time

• Progress_bar:turn on/off console progress bar during test run show progress bar

• Console_logging:turn on/off Logg Whether ing to stdout output to stdout

• Xml_report:turn on/off XML/JTL report generates XML reports.

• Results_database:database connection string (optional) to save the resulting database connection string (optional)

• post_run_s Cript:hook to call a script at Test completion (optional) the aftercare script (optional)

For more information see: http://testutils.org/multi-mechanize/configfile.html

Script writing

Writing in Python, the test script simulates a virtual user's request for a website/service/API, the script defines the user's transaction, and more on the Scripting Manual: Http://testutils.org/multi-mechanize/scripts.html#scripts-label.

Each script must implement a transaction () class. This class must implement a run () method. The basic test script structure is as follows:

Class Transaction (object): Def run (self): # does something here return

Transaction () is instantiated once during operation, and the run () method is called repeatedly:

Class Transaction (object): Def __init__ (self): # does Per-user user Setup here # The gets called once on U Ser creation return def run (self): # does user actions here # gets called repeatedly ret Urn

Class Transaction (object): Def __init__ (self): Pass def Run (self): BR = mechanize.  Browser () br.set_handle_robots (False) resp = br.open (' http://192.168.4.13/env.htm ') assert (Resp.code = =), ' Bad response:http%s '% resp.codes assert (' service name ' in Resp.get_data ())

below, rewrite the script with the Httplib library and add the timer. With timers, you can analyze the time-consuming aspects of each step.

Import httplibimport timeclass transaction (object):     def run (self):         conn = httplib. Httpconnection (' 192.168.4.13 ')         start = time.time ()         conn.request (' GET ',  '/env.htm ')          request_time = time.time ()          Resp = conn.getresponse ()         response_time =  Time.time ()         conn.close ()          transfer_time = time.time ()         self.custom_ timers[' Request sent '] = request_time - start         self.custom_timers[' response received '] =&Nbsp;response_time - start        self.custom_timers[' Content transferred '] = transfer_time - start         assert  (resp.status == 200),  ' bad response: http %s '  %  resp.statusif __name__ ==  ' __main__ ':    trans =  Transaction ()     trans.run ()     for timer in  (' request  sent ',  ' response received ',  ' content transferred '):         print  '%s: %.5f secs '  %  (Timer, trans.custom_timers[timer])

There are more instances of Http://testutils.org/multi-mechanize/scripts.html#scripts-label.

Resources

• https://pypi.python.org/pypi/multi-mechanize/1.2.0

• Http://testutils.org/multi-mechanize

• Author Blog: http://my.oschina.net/u/1433482

• Type: Finishing

Python Module Introduction-multi-mechanize Universal Performance testing Tool