PyMunin - Python Multigraph Munin Plugins
Python Module for developing Munin Multigraph Monitoring Plugins.
More detailed documentation for the project and sample graphs for plugins are published in the PyMunin Project Web Page
Regular Munin Plugins employ one-plugin one-graph logic and require the execution of a script for data retrieval for each graph. Multigraph plugins permit retrieval of data for multiple graphs in one execution run (one-plugin many-graphs), reducing the processing time and delay for the fetch cycle significantly. More information on Multigraph Plugins can be found in the Munin Wiki:
The plugins consist of the following components:
- The pymunin module (./pymunin) implements the base classes for developing Munin plugins.
- The plugin logic is implemented in the plugin scripts in ./pymunin/plugins.
- The actual data retrieval logic is separated from the plugins to facilitate code reuse. Individual modules in the directory ./pysysinfo implement classes for getting the monitoring data and returning them in dictionary objects. The separation of the data retrieval logic should facilitate the use of the same code in other monitoring solutions.
Although the solution is focused on implementing Multigraph Plugins the module also supports simple single graph plugins.
Multigraph Monitoring Plugins for the following applications are already included:
- Apache Tomcat
- Apache Web Server
- Asterisk Telephony Server
- Disk Usage
- Disk I/O
- FreeSWITCH Soft Switch
- Lighttpd Web Server
- MySQL Database
- Network Interface Traffic and Errors
- Network Connection Stats (netstat)
- Nginx Web Server
- NTP - Time Server
- PHP APC - PHP Cache
- PHP FPM (FastCGI Process Manager)
- PostgreSQL Database
- Processes and Threads
- System Resources (Load, CPU, Memory, Processes, Interrupts, Paging, Swapping, etc.)
- Sangoma Wanpipe Telephony Interfaces
- Varnish Cache Web Application Accelerator
Classes for retrieving stats are available, but no plugins have been developed yet for the following:
- Squid Web Proxy
The documentation for the project and sample graphs for plugins will be published in the PyMunin Project Web Page
PyMunin is copyrighted free software made available under the terms of the GPL License Version 3 or later.
See the COPYING file that acompanies the code for full licensing information.
PyMunin has been developed by aouyar (Ali Onur Uyar).
Some of the people that have knowingly or unknowingly contributed with the development are:
- Initial packaging of the code was done by mlavin (Mark Lavin). PyMunin is installable pip / easy_install thanks to Mark. :-)
- The initial design of the solution was inspired by python-munin by samuel (Samuel Stauffer).
- Many plugins were inspired by existing _Munin Plugins_developed by other people. (Before developing any plugins, I always try to check existing solutions.)
I hope that more people will be using PyMunin for developing plugins in the future.
The easiest way to install the code is to use pip:
- Install the newest version from PyPI: pip install PyMunin
- Install the latest development versión: pip install git+https://github.com/aouyar/PyMunin.git#egg=PyMunin
The other option is to download and uncompress the code and execute the included setup.py script for installation: ./setup.py install
For detailed instructions on the installation process please check the project documentation at PyMunin Project Web Page.
On error plugins return short error messages by default. Plugin debugging must be enabled to return full trace for exceptions.
To enable plugin debugging in munin-run use the --pidebug option.
I would be happy to receive suggestions on improving the code for developing Munin Plugins. Alternatively you can use the Issues functionality of GitHub to document problems and to propose improvements. You can use the internal messaging system of GitHub or my e-mail address in case you prefer to contact me directly.
I hope that by sharing the code, the existing plugins will get more testing and receive improvements, and many more Multigraph plugins will be developed collaboratively.
I would be glad to receive some sample graphs from anyone using the plugins.