Torstein Molland

01/12/2023, 9:59 AM
Hi, We are having an issue where sometimes only two flows are executed concurrently and the remaining flows are stuck in the "pending" state. Our work queue (currently only one) is set up with unlimited concurrency and we are not enforcing any concurrency limits. We are running the flows in AWS using ECS. I noticed that sometimes our memory usage is high at about 80%. Could this be the reason the concurrency is sometimes limited? Thank you kindly for your time 😊

Christopher Boyd

01/12/2023, 3:23 PM
Hi Torstein, without details and logs, it’s mostly just guessing, but it’s certainly a possibility. The job can be submitted to the infrastructure, but due to lack of resources the pod can’t be scheduled to a node, so it can’t actually start and go to “RUNNING”