breaking things would be an interesting topic here

[matthewfischer]

One good way to learn about how bosh handles failure would be to have a section on breaking things. Although some of this is covered in bits in pieces it would solidify the relationships between bosh/monit/diego? etc.

For example:

1. Delete the VM hosting zookeeper - check out bosh events and watch the AWS console
2. ssh to the box and kill the zookeeper process - who brings it back? (a: monit)
3. Detach and or delete a disk.
   etc

[aussielunix]

There is an ops file that sets up turbulence on your director and then exposes an API for injecting 'chaos' into your system.