Refinement of ResourceExhausted reasons

Is your feature request related to a problem? Please describe.
Currently, there are three ResourceExhausted errors with the RESOURCE_EXHAUSTED_CAUSE_BUSY_WORKFLOW reason.

1. ErrWorkflowClosing, in my opinion, does not necessarily mean exhaustion.
   https://github.com/temporalio/temporal/blob/aec2c100b6fae91a53f0f4ea3befe8ea683f0d9c/service/history/consts/const.go#L71-L72

2. ErrConsistentQueryBufferExceeded indicates that there are too many queries but BUSY_WORKFLOW might be too ambiguous in this case.
   https://github.com/temporalio/temporal/blob/aec2c100b6fae91a53f0f4ea3befe8ea683f0d9c/service/history/consts/const.go#L77-L78

3. ErrResourceExhaustedBusyWorkflow explicitly includes "busy workflow" in its name. So, this error looks fine :) However, the comment indicate that this error should not be retried, but resource exhaustion usually means "retry later, please".
   https://github.com/temporalio/temporal/blob/aec2c100b6fae91a53f0f4ea3befe8ea683f0d9c/service/history/consts/const.go#L101-L102

These errors are indistinguishable in metrics, so I cannot exclude ErrWorkflowClosing (which is valid in my workload) from alerts.

Describe the solution you'd like
More granular reasons for service_errors_resource_exhausted.