other resources:

  http://vitobotta.com/resque-automatically-kill-stuck-workers-retry-failed-jobs/#sthash.oQsaNeb5.dpbs
  http://stackoverflow.com/questions/10757758/find-out-if-a-resque-job-is-still-running-and-kill-it-if-its-stuck

heartbeat?

redis scenario
  different processes (cant share state in Resque memory etc.)
  different boxes! (redis is shared state)
  if this is included in servers, you have clusters os unicorn/mongrel etc, 
    ensure it works across multiple running instances of this lib?

  Thing.setup_checker_thread
  Thing.enqueue_repeating_refresh_job

reproduce the error and integrate test like that

## TODOS

verify @config options, raise if no handler, etc.

- use locks to only notify/trigger once in case this is used in a server-cluster type environment adn you only want the handler triggered once

- resque in stuck (t state process) SIGSTOP sticks up the checker thread?
  or if not running

- ensure it only runs from the server box and not the resque box??
  (the deploy restarts the server but not resque workers)

add a 'resque_stuck_queue/tasks' bit? See tres eg
require 'resque/stuck_queue' instead?