Add metrics for the runner for picking up jobs

[lispyclouds]

Since the docker runner is concurrent and the numbers of runs is limited by the machine resources, having more and more jobs in the queue would cause the runner to pick them all up and be overwhelmed.

• Have a way for the runner to decide whether it can pick up more jobs
• Use metrics like CPU and/or RAM usage to check this maybe?
• If the runner isn't capable, it should put the message back for other potential runner(s) to pick it up