Create a RunQueue that manages the local compute resource - limiting the number of active processes.
- Group Reviewers
Couple questions, but overall I think this looks good! Will let max and schrockn take a look also.
maybe a helper function that takes in the different queue configs and runs the test?
Took a while to confirm the difference between test_local_compute_queue and test_delegating_queue
do you think DelegatingRunQueue will be a common case? Is this mostly to set up different priority queues?
should we do a check here for non-null PipelineRun to guarantee that we don't accidentally stop the run queue processing?
the test bodies are a bit different 
not sure how common - but ya so you could set up different queues for different resource requirements ie dagster/queue/needs_gpu