PyMoodss

Extension library for moodss written in python

What is PyMoodss?

PyMoodss is an python extension library for the moodss server monitoring system developed by Jean-Luc Fontaine in the tcl programming language. While moodss originally was designed to monitor one server or a limited number of servers, PyMoodss enhances moodss to monitor serveral servers or server-clusters.

Architecture

PyMoodss delivers a flexible framework of four base-classes written in python:

• Report parsing class (filtering of mission results)
• Mission class (wrapper for the tasks e.g. "ssh root@xx.org -l ps")
• Server class (contains all necessary server informations domains, ports,..)
• Servers class (Container for all Server, parsing of config files, analysing dhcpd.conf files)

intended to be subclassed to form moodss-modules. Only subclassing of the two first classes is essential to form a new module. Modules written using PyMoodss serve the tclpython-interface created by jean-luc. Since the library is multithreading this enables PyMoodss-Modules to monitor a lager number of servers "in parallel".

Avaible moodss Modules based on PyMoodss

Utilizing Pymoodss several moodss modules were developed, covering

• mod_backhand (apache clustering module)
• df (fill state of partitions)
• psgrep (checking the states of a set of processes on serveral machines)
• mping (ping statistic for a lot of servers)
• mdstat (state of RAID-devices)

extending the set of TCL moodss modules avaible for moodss.

Download

The PyMoodss package is avaible on http://www.sourceforge.net/projects/pymoodss/.

Configuration of the PyMoodss modules

The configuration of the PyMoodss modules is done in an simple unified way. E.g. To configure psgrep to monitor the processes httpd, sshd and ntpd on the machines alpha, beta, delta the psgrep configuration file reads:
[alpha, beta, delta] [httpd, sshd, ntpd]
as a cross product rule. Doing ping on the same machines the ping modules config file reads:
[alpha, beta, delta]

Maximal risk filtering

PyMoodss introduces a very desirable feature to moodss: One Trigger on the same property of a large number of servers.
[If you are monitoring 5 Servers - it's no problem to define 5 triggers (via the GUI): "if the fill of partition x is over 70% send me an warning email". If you are running 50 servers this becomes a stupid task]
PyMoodss solves this problem by defining filter rules that generate e.g. one moodss result line containing the partition that has actually the largest fill of all servers over the whole server-cluster. Therefore only one trigger has to be set on this special (Maximal Risk) line.

Screenshots

A screenshot of moodss with PyMoodss modules running is avaible.

Author

PyMoodss was developed by my person at inqbus it-consulting to monitor the large german student portal Unister (12 Server-cluster) using the great open source software moodss. Now we like to give back our drop of open source to the ocean we have used while our developement.