Create GET endpoints for ClusterRunner logs

Currently in order to view the logs for: clusterrunner_master, clusterrunner_slave, clusterrunner_deploy, etc, we need to ssh into the particular machine on which those commands ran and open the files in /home/jenkins/.clusterrunner/log/clusterrunner*.

This is a pretty big pain, and makes it difficult to automate any log aggregation process.

Chad proposed that we create a GET endpoint that would return a JSON array with each log entry being its own JSON array element.

This endpoint would need to accept arguments that at least included:

which log was requested (slave/master/deploy/stop)
how many entries to return (as the number of entries could be huge)
(optional) filter by error, warning, information, notice