One day I had this fantasy of starting a certification service for operations. The certification assessment would consist of a colleague and I turning up at the corporate data center and setting about critical production servers with a baseball bat, a chainsaw, and a water pistol. The assessment would be based on how long it would take for the operations team to get all the applications up and running again.

This may be a daft fantasy, but there's a nugget of wisdom here. While you should forego the baseball bats, it is a good idea to virtually burn down your servers at regular intervals. A server should be like a phoenix, regularly rising from the ashes.[1]

The primary advantage of using phoenix servers is to avoid configuration drift: ad hoc changes to a systems configuration that go unrecorded. Drift is the name of a street that leads to SnowflakeServers, and you don't want to go there without a big plough.

One way to combat drift is to use software that automatically re-syncs servers with a known baseline. Tools like Puppet and Chef have facilities to do this, automatically re-applying their defined configuration. [2] The limitation is that re-applying configuration like this can only spot drift in areas that you've defined that the tools control. Configuration drift that occurs outside those areas doesn't get fixed. Since phoenixes start from scratch, however, they will pick up any drift from the source configuration.

This doesn't mean that re-applying configuration isn't useful since it's usually faster and less disruptive than burning down a server. But it's valuable to use both strategies to fight away the snowflakes.

Further Reading

Netflix has a chaos monkey that randomly burns down servers in order to test that their system is resilient.