Jobs on RabbitMQ queues are limited to only one executing job at a time per queue for each server

Queue consumers - one of which is created for each Hangfire queue - are configured with a `prefetchCount` setting of '1':

https://github.com/HangfireIO/Hangfire.SqlServer.RabbitMq/blob/e72d8ee7bab48b751f37b4e37b7ddf0f32523faf/src/Hangfire.SqlServer.RabbitMq/RabbitMqJobQueue.cs#L111

The result of that is that each consumer will only be allowed to have one message (job) unacked - which they are whilst being processed. Even if more workers are configured and available on a given server, they are all made to wait for one job to be completed until another one can be started.