It doesn't matter how much care and attention you apply to your platform; something can and will inevitably go wrong when you least expect it. For example, a server will fail, a process might start looping, a network switch will decide it doesn't want to be a network switch anymore, the SAN will decide it likes to function in single-user mode, the latest security patch will cause issues in the software platform, or someone will decide to hack you because you're a nice big target. As the saying goes, you should always expect the unexpected.
Most businesses will have some sort of business contingency plan in place to cater for the unexpected, but there's a strong possibility that they don't try to purposely force the issue, at least not to the extent that something bad actually happens—they just hope nothing bad will ever...