Name                    Last modified       Size  Description

 Parent Directory        05-Jan-2009 18:48      -  
 ANNOUNCE                14-Sep-2004 21:20     1k  
 COPYING                 12-Sep-2004 00:33    18k  
 ChangeLog               14-Sep-2004 21:21     1k  
 DESCRIPTION             15-Sep-2004 01:39     1k  
 Makefile                15-Sep-2004 01:40     3k  
 Makefile.tino           13-Sep-2004 04:20     1k  
 PROTOCOL                13-Sep-2004 03:14     4k  
 README                  13-Sep-2004 04:39     2k  
 VERSION                 14-Sep-2004 21:13     1k  
 mon.c                   15-Sep-2004 01:39    20k  
 mondump.c               13-Sep-2004 03:21     6k  
 tino                                           -  

$Header: /CVSROOT/mon/README,v 1.3 2004/09/13 02:39:51 tino Exp $

You can find this at URL: http://www.scylla-charybdis.com/tool.php

This is a lightweight system monitor for case you have no time.

Just compile and do
	nohup ./mon [interval] >>mon.out 2>mon.err &
Then logoff.

Mon then gathers the system valid data for you each interval seconds,
or 5 seconds if interval is missing.

If you have the time later, you can run the monitoring information in
file mon.out through some evaluation, like rrdtool.


Additional programs:

	mondump < mon.out
or
	./mon | ./mondump

This makes a testdump of the monitoring data just to proove, that
the concept is OK.


Programs to come soon (I hope):

	mon2rrd [offset] < mon.out
or
	./mon | ./mon2rrd

Creates a RRDB file tree for data found in mon.out in the current
directory.  This is somewhat stupid and magic, as it assumes, mon.out
has 5s gathering interval.

Offset (default 0) is the time offset (in seconds) of the system to
UTC.  So the timestamps in mon.out + offset must give the correct UTC
time.



Future functionality:

mon [options] host[:port] [maxhist]

This will send it's data to a logging server at host:port (default
2606) using UDP.  It will use a sliding window of maxhist K gather
points (default 128K, which is a little bit more than one week for a 5
seconds interval) to store the data in case the host is down or
packets are lost.  The host is then able to retrieve this information
in case the packets were lost.

mond ip[:port] [list of allowed IPs or networks]

The protocol uses some challenge/response to safeguard against
spoofing, such that both sides are protected against blind spoofing.
Except from this everything is clear text (use a VPN tunnel to protect
the communication if you wish).  Additionally, the protocol is
stateless, so you can restart the logging server at any time without
loosing any information.  If you restart the client side, then you
only loose the information which was not yet transmitted successfully.


TODOs:

- Monitor system files and send all changes to the other side.

- Monitor syslog (gives a reliable forward to the other side).

- Compression and Encryption support

-Tino, webmaster@scylla-charybdis.com
$Log: README,v $
Revision 1.3  2004/09/13 02:39:51  tino
bugfix for wrong sizeof parameter in memcpy() of index_new

Revision 1.2  2004/09/13 01:21:17  tino
preparing dist

Revision 1.1  2004/09/13 01:14:43  tino
first version without network monitoring