Yes. But one thing i would like to do is to make sure the that the worker pulling from the fair shared queue is executing the task as if it was coming from temporal as an activity handler to minimize the transition when temporal support it. Basically from the handler’s perspective nothing really changed, we’ve simply created a transport protocol through another queue (man in the middle).
As for Hatchet, i think their formulation is a bit confusing. They explain it better in Round Robin section. In the end they are essentially splitting the queue into multiple sub-queues internally, like per tenant, and round robin between them. As far as i know they don’t support multi-level, like per tenant per task/activity type, which i would like, at least not yet.