This is collection of cricket modules which use ssh to acquire data from remote hosts with minimal or no installation on monitored host. non-interactive login on remote host and execute different command according to identity file. Most documentation is included on top of perl file. There are also example Default and target files in respective directories in distribution. You will have to edit Default files if your installation is not in /data/cricket. maillog ------ Install parse_maillog.pl on target host and ssh to get data in cricket. You will have to change some parameters in script: $log to point to your mail.log location if you don't want to give log as fist parameter in command line (recommended) $domain to your local domain (so that you can separate incoming and outgoing e-mail in graph) $delta (by default /var/tmp), location of temp file which will store current offset in mail.log -- you should create that file under user which will run parse_maillog.pl or you risk race bug by malicious user iptraf ------ This module is based on work of Eric Wong available at http://www.taedium.net/rrd-iptraf/. You will have to install parse_iptraf.pl on target host and edit: @PORTS array if you want to monitor something else too (have in mind that you will have to edit iptraf/Defaults also if you change @PORTS!) $log - only if you don't intend to give parse_iptraf.pl log as parameter (I recommend that) $delta - to move offset file somewhere else than /var/tmp You will also have to setup iptraf to monitor traffic in 5 minute intervals. Do that from courses interface and start iptraf with # iptraf -s eth0 -B For detailed information about iptraf setup see http://www.taedium.net/rrd-iptraf/. Since you are NOT going to use rrdtool directly (cricket will call it for you) you can ignore rest of instructions. bind9 ----- You should configure your bind9 on target host as described in README.bind9stat. Then configure parameters on top of script or use command line parameters. df -- You don't need to install anything on destination host (other than authorized_keys). I recommend that you use command 'df -Pl' to display df output in format which parse_df.pl understands and to display only local file systems (tested on Linux, HP-UX and Digital Unix). This module will use CPAN perl module Cache::FileCache if available to provide caching (which means that it will connect to each host just once to fetch all disk information). Since you have to install that module on your monitoring station (on which you run cricket) and NOT on target host, I recommend that you do so. You can also use $ parse_df.pl "ssh -i ~cricket/.ssh/df target.host" --config to create target file (with all local partition on target.host) for you. ntp --- This monitor does not require installation on target host. Just run $ ./parse_ntpq.pl ntp.domain --config to create dump target file for ntp server 'ntp.domain'. Most modules will create skip-overview var in target file, so that script generate_overview.pl will show just summary graphs. You might want to remove that from target files if you don't want to use generate_overview.pl generate_overview.pl -------------------- This script generates html pages with daily, weekly, monthly and yearly overview of all services and html page per service. It's currently in state of flux. It works, but it's not well documented and needs additional work (and editing of paths in it to work for you). If you use it, remember to change paths on top of file. For distribution site see http://www.rot13.org/~dpavlin/sysadm.html (there is also ChangeLog and CVS repository there).