Produce problems for yourself

Netflix has moving their application to the cloud, leveraging the hardwards in Amazon EC2.
They adopt an architecture call micro service architecture, which breaks normal components in to microservices. These microservices work together to achieve a task.

Microservice could be failed - maybe due to network congestion or simply unavailable. To proactively deal with failure of microservices, they make use of Chaos Monkey in Amazon EC2 that could purposely turn off the microservices. This helps them to see their resilients against failure.

I feel this is a good approach. Rather than waiting for problems, PRODUCE problems for yourself. This allows yourself to have time for fixing or preparing it. Nicely done!