A case for an automated web server monitoring service

I run a popular travel-related website with over 1,500 unique visitors each day. As the site is essentially a Discussion Board (using the popular vBulletin software) it is extremely CPU intensive. To ensure good response times, I recently moved the site from a shared-hosting platform to a Virtual Private Server (VPS).

A VPS is effectively your own server as you have full access to the root directory. However, as you do share the physical hardware with other websites, there is the added benefit of additional CPU during brief “peak” periods.

Things went very smoothly for the first few months. The VPS provided a fast and reliable service. During the occasional very busy periods, the shared CPU kicked-in offering outstanding speed even during these peak times.

What I didn’t realize was that my VPS provider had installed an automated VPS management system which managed the shared CPU usage. If shared CPU usage exceeded predetermined levels, the VPS management system simply closes down the relevant server as explained on websites such as hardware-performance.com

And that’s exactly what happened to me! To make matters worse, I was overseas at the time and unable to manually monitor the website. There was an article about the website in a local newspaper which generated a lot of new visitors to the site. The additional load on the server, caused the shared CPU to kick-in, when in turn trigged the VPS Management system to shut down my site. I only became aware of the problem when I returned. The users of the site were upset and I suffered significant revenue loss.

This problem could have easily been avoided had I used an automated web server monitoring service. These services try to access the server at regular intervals. If it doesn’t get a response, it alerts you by email or SMS.