One of the sites that I work on, EarthTrends, used to have a stability problem. It was running on a Windows server using Coldfusion and a SQL Server back-end, and we were intermittently having problems with Coldfusion crashing the server. After many months of troubleshooting and tweaking, we decided to switch to a LAMP platform - a decision that was partially because of the stability issues and partially because we wanted to take advantage of open-source technologies. During that period, I wrote a custom monitoring script in Coldfusion, that checked the website for availability every 5 minutes and had an escalation procedure for notification emails and text messages. This worked like a charm for over a year, then we de-commissioned the servers and switched to a LAMP platform. My script disappeared into oblivion, but we never had any problems in our LAMP environment.
That was until this week. We had our first Apache failure that was related to a configuration change - turning KeepAlives on. This may be due to the heavy volume of robots that crawl the site, but the result was an un-responsive Apache. We had no monitoring script to catch the problem and it went about an hour un-noticed.
There are many different monitoring systems out there - from the simple to the full-featured - but what I needed was something that was simple and required no setup. Enter Montastic. Montastic is a free web service that will monitoring an unlimited number of sites every five minutes and provide notification to multiple email addresses when the site goes down. Right now I have it set to notify my work email and send a text message to my cell phone, a feature that I have with Cingular (just send an email to @cingularme.com).
The problem on EarthTrends is now fixed, and I can rest easy knowing that I will be notified if there are problems. Obviously, this is a super-basic solution and there are many more sophisticated ones out there. But, it is a good one if you just want the peace of mind that your website is up and running.