[Stuart]

One of the sites I maintain at work had to be taken off line yesterday when the software running a service it relies on suddenly decided to commit suicide at some point between 12am and 9am yesterday morning. After several hours, my colleagues managed to get the service back online, and the site is back. That's not the good news.

After being told by the service's original designer (it was custom written in Java) that it was impossible, I've managed to write a little bit of code that, should this service fail again, takes the site off line and emails me with the exact details of the error given by the service.

So, while it won't stop the service malfunctioning in the future, it should at least give me some advance warning when the service does malfunction.

Admittedly, the code is only working on our test site at the moment (it will be implemented on the live site along with a few other changes when we have tested them properly), and while the code functioned correctly when I was able to engineer failures of the service (the test site uses a test version of the service), but there is no real way to be 100% sure what will happen with an actual failure.