ANNOUNCE: moodss-21.0 and moomps-5.4
jfon...@free.fr
By finding the best of powerful statistical models, using sophisticated
methods such as ARIMA (AutoRegressive Integrated Moving Average) and
artificial neural networks, moodss can now predict the future behavior
of data cells, from their history recorded in a SQL database. The new
predictor tool, obviously ideal for capacity planning, will also allow,
in upcoming releases, a system administrator to receive emails such as
"on server foo.bar.com, the disk sdb is likely to become full in 3
weeks".
You may find the specific documentation of this new feature, including
screenshots, at http://jfontain.free.fr/statistics.htm. This new
release also includes common sample time series to practice on, so if
you are a statistician, please try it: your comments will be most
welcome.
### CHANGES ###
--- moodss 21.0 and moomps 5.4 ---
- in moodss GUI:
- implemented new predictor tool for future behavior prediction
based on past data cell history from moodss SQL database (not
available for Windows)
- implemented interface to R statistical calculation engine:
- glue procedures for ARIMA functions
- native code and glue procedures for neural net functions
- multiple concurrent R processes and priority management
- added Tool/Predictor menu
- in preferences, added Tool/Predictor page
- included common time series samples for user training on predictor
- added HTML document also used in moodss application help
interface, complete with examples, screenshots, manual and
statistics methods documentation
- when loading another instance of a module from the database, the
View/Objects menu was not properly updated
- View/Trace menu is now also usable in database mode
- allowed loading dashboard from standard input (allows
pre-processing) in command line using the -f(--file) option with
"-" as argument value
- usage was not printed in case of missing or bad option argument in
command line
- in case of database error, View/Trace menu did not show that trace
window was opened
- increased main window size so that a new predictor tool window
fits
- in database layer, for ODBC connections, let the ODBC layer handle
locks via transactions (except for MySQL), which solves SQL Server
problem on new module instances
- updated myvars, myhealth and mystatus modules for MySQL 5.0 support
- in moodss rpm specification, required R package (available in Fedora
Extras)
- includes md5, uri, base64, smtp and mime source from new tcllib 1.8
release
- in standalone binary, upgraded Tcl/Tk to 8.4.12 from 8.4.11
- in standalone binary and Windows distribution, included SQLite 3.2.8
### README ###
This is moodss version 21.0 and moomps version 5.4, powerful modular
monitoring applications.
For Unix Review, moodss is "a must-have application for today's
network and systems administrators", and for Eric S. Raymond, in "The
Art of UNIX Programming" book: "the code is polished, mature, and
considered an exemplar in the Tcl community". For Joe Barr, at
NewsForge: "I downloaded the moodss tarball from the website,
decompressed it, and started it up. It's that easy. The main window is
deceptively simple. Great power lurks just below the surface of that
mild exterior".
Moodss is a graphical application, which, in real-time mode, displays
data processed by any number of dynamically loadable modules. Various
data tools (graphs, pie charts, formula builders, thresholds manager,
...) are used to build complete dashboards, very easily by drag'n'drop,
to monitor a single server up to a whole information system. Any
displayed data can also be archived in a SQL database, which moodss
can use for post-analysis and presentations, or even capacity planning
by predicting the future, using sophisticated statistical methods and
artificial neural networks.
Moodss companion daemon, moomps, works similarly around the clock, and
even allows distributed monitoring to feed remote moodss stations.
Modules, the link between the moodss and moomps cores and the
monitored data, can be easily created (in Tcl, Perl, Python, HTTP, C,
...).
Many modules are provided, such as a comprehensive set for Linux
system monitoring, MySQL, network, SNMP, Nagios compatibility, Python
and Perl modules examples. For example, thoroughly monitor a dynamic
web server on a single dashboard with graphs, using the Apache, MySQL,
ODBC, cpustats, memstats, ... modules. If you have replicated servers,
dynamically add them to your view, even load the snmp module on the
fly and let your imagination take over...
There are currently about 100 usable modules for moodss (counting the
Nagios plugins)
Moodss is multi-lingual: English, Japanese and French are
supported. Help with other languages is very warmly welcomed.
###
You may find it now at the following locations:
http://download.sourceforge.net/moodss/moodss-21.0.tar.bz2
http://jfontain.free.fr/moodss-21.0.zip
http://jfontain.free.fr/moodss-21.0.i386.tar.bz2
(note: rpms also available in Fedora Extras repository)
http://jfontain.free.fr/moodss-21.0-1.i386.rpm
http://jfontain.free.fr/moodss-21.0-1.spec
http://download.sourceforge.net/moodss/moomps-5.4.tar.bz2
http://jfontain.free.fr/moomps-5.4-1.noarch.rpm
http://jfontain.free.fr/moomps-5.4-1.spec