Many organizations
have over used on hardware without realizing that. About a year ago I looked at
proper server monitoring equipment, and of course being Linux servers I desired
to monitor I identified some very capable totally free open source solutions.
Initial is munin, which runs more than a web interface, but has standalone
server therefore you can monitor Apache, along with mail companies and any
other services you may have to keep an eye about. It also emails each time a
service restarts or drops dead. Using an email server like postfix you are able
to send a text concept alert by piping the email content through a Perl
software and sending SMS (this can require an online service which includes an
API for Perl, We have used the o2sms package deal, which in Ireland covers
every one of the cell networks). Also it really is posssible to monitor server
with a WAP gateway for munin allowing us to check and also restart services and
server status when an net connection is not otherwise accessible.
Another
nice piece regarding software is monit, which generates reports on almost every
aspects a sys admin would have to know about a method, CPU, disk usage, jobs,
mail queues and a bunch of other useful tid-bits plus a full historical
graphing for some too... this can easily recognize bottlenecks or under use of
server resources. In particular it pays to for monitoring mail lines and mysql
queries which is often stress points every so often.
Plenty
of details about setting these up and getting started is found on http://www.woodstone.nu/salive