In what order are requests are assigned?
Here's how help desk requests in an automatic queue are assigned to agents.
The application calculates the priority of help desk requests based on severity and waiting time. Help desk requests with the highest severity are assigned first. If there are multiple requests with the same severity, requests that have been in the queue for the longest time are assigned first.
It's possible that you may see exceptions to this rule when you're testing with a small number of agents. That's because the application is designed to work with millions of help desk requests and thousands of agents per day. And to do that, it has massive parallel processing power.
There are multiple servers in your production environment. Each server tries to assign help desk requests , competing with other servers to assign all the work among all available agents. Requests are assigned on multiple servers in parallel, otherwise the assignment tends to lag behind the incoming work.
Suppose your production environment has 15 servers. The application would act as if it's assigning help desk requests in batches of 15. The requests from 1 to 15 would be assigned before the requests 16 to 30. But the order within each batch isn't guaranteed. The first server that finds an available agent makes the assignment.