I just pushed a example on how to create a Slim framework to github. The idea is that whenever you create a webservice with the Slim framework (which is really simple) you rarely test for failure, the ChaosMonkey class will help you to do just that. When initialized with the AbsoluteChaos plugin it will randomly kill the webservice with exceptions, garbage to the output, or just run the service for you without failure at all.

His plugin does a lot of things right now, but it's easy to extend with your own failure types - like his suggested "networkSleep" or something that could kill the connection to MySQL. He includes a code snippet in the post of how to hook Slim and ChaosMonkey together for some testing fun.