When a cluster is under heavy load it is helpful to the user if they get some indication that their request will take longer (or not complete). By sending an alert back the first time a request is not able to get any locks (and then not sending any more alerts due to that condition for the rest of the request), the user will have some indication of what is happening.
When a cluster is under heavy load it is helpful to the user if they get some indication that their request will take longer (or not complete). By sending an alert back the first time a request is not able to get any locks (and then not sending any more alerts due to that condition for the rest of the request), the user will have some indication of what is happening.