Hey, is there a way to configure the concurrency inside an job? The docs only talk about the concurrency jobs...

you can limit the number of concurrent steps that are running across all ops in a job when using the multiprocess_executor. I think this can be achieved on a more fine-grained level using the celery executors as well (listed here ), but I'm not as familiar with those

Is it also possible when running it in ECS?

yes, your job running in an ECS container (launched with EcsRunLauncher) can be configured to use a variety of executors. The multiprocess executor is one of those possible configurations (also the default). The docs I linked in my first reply shows how you can configure it in your run config, and also notes about default parallelism and execution priority tagging

Thanks, I will take a look