In Orchestrator 2012, you have runbook servers that execute your runbooks. Those runbook servers have throttling limits, that limit the maximum amount of concurrent runbooks that can run on a runbook server. By default, this value is 50, but it can be increased/decreased, depending on your environment. If your primary runbook server is executing 50 concurrent runbooks, and a 51st runbook is started, the runbook will fail over to a secondary runbook server. If the secondary runbook server is not available, then the runbook will queue up until a slot becomes free on a runbook server.
This leads to an interesting question when you are dealing with monitoring runbooks. A monitoring runbook sits and monitors until its trigger condition is met. Once the trigger condition is met, a new instance of the runbook is spawned to continue running, while the original runbook completes its execution. But what happens if a runbook server is currently at its throttling limit (for example, 50 concurrent runbooks), and one of the monitoring runbooks triggers? Does it still spawn another process, even though that would make 51 runbooks running on the server? Does it queue up?