I’m trying to get a sense for the temporal “Workers” reported in the UI vs the number of nodes running workers. Currently I have 100 pods running my workers but the UI reports roughly 40 “Workers” at any given time. These are very busy workers. Just trying to get a sense for the expected behavior here?
These are very busy workers.
If the workers do not have slots for activity or workflow tasks, they will stop polling (until one slot becomes available). Could you check the SDK metric worker_task_slots_available
Antonio