Python-time Intro

A transaction statistics profiler.

the following project is an profiler framework was writen for handle, manage, and get transaction response time statistics. it also extends profilehooks, and expose some additional capabilities.

some of the advantages is the ability to publish stats to time series db.

screen shot

Advantages:

• measure response time.
• measure network IO correlated to response time.
• publish the results to time series db (influxDB).

Quick Start

Requirements

python pkgs:
this programs require the following python dependencies:
influxdb, concurrent.futures, psutils


setup components:
- influxdb
- grafana

Usage And Examples

setup

1. clone the project to where you want to monitor the python code.

   git clone https://github.com/mrsiano/python-time.git
   cd python-time
   # make sure your influxdb running.
   

2. export python-time

   export PYTHONPATH=<gitdir>/python-time/
   

3. edit the configuration file see config.cfg

   vim config.cfg
   make sure to set the influxdb: server \ port.
   make sure to specify the network device: net_device = en4
   

4. decorate your codes

   ./decorating_app.sh <the python code you want to monitor>
   

more standalone examples.

1. decorator usage:
    first the module should be imported e.g:
    from transResponseTime import measure_time
    
    then simply decorate functions
    
    @measure_time()
    def a_test_method(sec=0.2):
        time.sleep(sec)
        
     is it possible to use the profilehooks syntax e.g:
     @measure_time(immediate=True)
        
2. Plain usage:
    first the module should be imported e.g:
    import transResponseTime
    ...
    transResponseTime.measure('plain_usage_exam', time.sleep, 0.5)

Analysis

once your test completed and your influxdb collects the results, import the grafana template into your grafana server.

1. first add to grafana your influxdb server.
2. make sure you change the db name in the json template file before importing.