Flower exposes Prometheus metrics, which is incredibly useful.
Idea: expose a 'worker concurrency' metric. As Flower already exposes the current amount of running tasks (flower_worker_number_of_currently_executing_tasks), exposing the max amount of handleable tasks would allow for monitoring whether a worker is too busy.
While the same could probably done with flower_worker_prefetched_tasks, having a fixed concurrency metric might be a bit easier to graph.
Flower exposes Prometheus metrics, which is incredibly useful.
Idea: expose a 'worker concurrency' metric. As Flower already exposes the current amount of running tasks (
flower_worker_number_of_currently_executing_tasks), exposing the max amount of handleable tasks would allow for monitoring whether a worker is too busy.While the same could probably done with
flower_worker_prefetched_tasks, having a fixed concurrency metric might be a bit easier to graph.