Add a catchall for unhandled failure cases

[mlehner616]

This adds a catchall to log very useful events that are not displayed anywhere else upon deployment failure.

For example:
This output isn't very helpful to developers in deploy output. There's no explanation on why it failed.

2018/04/17 12:40:59 PDT [ERROR] levant/deploy: deployment ee8fb7f8-4277-11e8-a099-17687cc3cd50 has status failed
2018/04/17 12:40:59 PDT [ERROR] levant/failure_inspector: alloc fee09c3a-4277-11e8-b4ac-4fe73209d2f7 incurred event killed because the task was successfully killed

This change would add the missed events that were not picked up by the case statement.

  2018/04/17 12:40:59 PDT [ERROR] levant/deploy: deployment ee8fb7f8-4277-11e8-a099-17687cc3cd50 has status failed
+ 2018/04/17 12:40:59 PDT [ERROR] levant/failure_inspector: alloc fee09c3a-4277-11e8-b4ac-4fe73209d2f7 logged for failure; event_type: received; message: task received by client
+ 2018/04/17 12:40:59 PDT [ERROR] levant/failure_inspector: alloc fee09c3a-4277-11e8-b4ac-4fe73209d2f7 logged for failure; event_type: task setup; message: building task directory
+ 2018/04/17 12:40:59 PDT [ERROR] levant/failure_inspector: alloc fee09c3a-4277-11e8-b4ac-4fe73209d2f7 logged for failure; event_type: template; message: missing: vault.read(secret/mysecret/key)
+ 2018/04/17 12:40:59 PDT [ERROR] levant/failure_inspector: alloc fee09c3a-4277-11e8-b4ac-4fe73209d2f7 logged for failure; event_type: alloc unhealthy; message: task not running by deadline
  2018/04/17 12:40:59 PDT [ERROR] levant/failure_inspector: alloc fee09c3a-4277-11e8-b4ac-4fe73209d2f7 incurred event killed because the task was successfully killed

[jrasell]

awesome @mlehner616 thanks for this