Ability to set maximum number of concurrent executions across my function host

Requirements:

[P0] In order to better allocate resources within a given host environment, we must have a setting which enables us to control how many concurrent executions are allowed to run at any given point for a given host.

[P2 ]We should be able to set the concurrency limit on different instances of a given host to a different value (in the case of host environments with different resource availability).

Additional thoughts:

For non-HTTP based events, this may be pretty straightforward. We may want to think through how this would affect HTTP based events. Would they be queued until a previous execution finished? If so, could we just give them priority, which would force the scale controller to create more instances to handle any queue messages, etc. that are piling up?

No files, tests, or entry points are named in the issue. Start by tracing how host settings and execution concurrency are configured, then resolve the behavior for non-HTTP and HTTP events, including per-instance limits and the definition of done.

Written by the indexing model from the issue text.