Some useful tools used for monitoring your site can aid you as the user to keep your site up, keeping your business alive.

Most companies will monitor your site for you and act on issues, however, having your own alerts sent to you via email or SMS maybe the way forward as it allows you to dig deeper into the issue as a user. Nagios (http://www.nagios.org/) Zabbix (http://www.zabbix.com/) are some of the in-house tools used by companies to monitor sites memory usage, filesystems, etc – but these usually require a server to be installed on a server fit for purpose, so for the average user thats not going to be possible. Using a third party company (such as UptimeRobot Pingdom Montastic) who offer the service online – some at cost, some free – a google search will provide a greater list of options.

Monitoring is the first step, but what if your site has issues? It might be worth taking preventative measures to keep your site up even when you’re having an issue. A good, free open source tool is http://mmonit.com/monit/ . You can use Monit to monitor daemon processes such as those started at boot from /etc/init.d/ for example apache mysql sshd send mail and mysql. What makes Monit different is that it can act on an error – for example if apache is not responding, restart the service, or if there are too many apache processes running like in a DOS attack. Monit can also check on files and directories and file systems checking for file size changes and time stamp changes which can be useful for making sure a partition on the server is not filled up – which can stop websites from running. Monit can also monitor network connections for main protocols like http and smtp. It can also be used to test programs or scripts at certain times similar to cron. Monit can be used for general monitoring such as load cpu usage and memory usage. This is a nice safety net for keeping sites online and a server contactable should there be an issue. This should not be treated as a fix more as a stop gap. Monit will then email you so you can analyse why, for example, apache is hitting this threshold – you may require a hardware upgrade or need to review code.