Hi, I tested DefaultRunCoordinator and QueuedRunCoordinator with our job and somehow QueuedRunCoordinator showed better throughput. Can you explain the reason??? There were multiple runs(200) and I think both run coordinators can run it concurrently. Is there any limit on concurrent runs on DefaultRunCoordinator ?

This is because the

QueuedRunCoordinator

dequeues runs using a threadpool. This was released in `1.1.7`: https://github.com/dagster-io/dagster/releases/tag/1.1.7

Hi, thank you for the answer!! But even if I reduced the number of dequeueNumWorkers and sensor:numWorkers to both 1, it shows better throughput than DefaultRunCoordinator. Actually, there is no difference between There is almost no difference between configuration A(dequeueNumWorkers 200, sensor:numWorkers 200), and configuration B(dequeueNumWorkers 1, sensor:numWorkers 1). Can you explain why it happens?? I confirmed that the configuration is correctly applied by checking GUI ->Run -> Queued -> View queed configuration