Darragh
06/25/2020, 11:36 AMLaura Lorenz (she/her)
06/25/2020, 2:14 PMDarragh
06/25/2020, 3:10 PMawslogs-stream-prefix
- the ID’s on those logs look like this prefect-flow-group/HUGE_UUID
, but the HUGE_UUID
doesn’t correspond to the ID of the Flow run itself. Is there way around that? It’s making log correlation a major PITALaura Lorenz (she/her)
06/25/2020, 3:41 PMDarragh
06/25/2020, 7:36 PMLaura Lorenz (she/her)
06/25/2020, 7:38 PMDarragh
06/30/2020, 11:12 AMSubmitted for execution: Task ARN: arn:aws:ecs:us-east-1:REDACTED:task/c07abccd-7dd4-4148-a55d-d682c4662b60
That sat for about 10 minutes in Submitted before I killed it - what happened then was the retry on the triggering task caught the “Failed” state, restarted the sub flow, and then it then started and completed successfully. My current guess on this is something to due with resource checks or API limits on AWS? But having said that, each trigger_subflow
mapped task waits for [ wait period ifor a trigger is a counter up to the number of flows multiplied by 5 seconds, so (0…20) * 5
] between each mapped task run before it actually triggers the flow, so they should be staggered.
It’s trying to create 20 instances of this flow at once, each running as a Fargate Task so it’s plausible that there’s an issue, but looking at it as Le Dumb Ass User there’s nothing to actually tell me that.
I’l tidy up the sample code I have and see if I can get it ready for a GitHub issue if we can’t figure it out 🙂Laura Lorenz (she/her)
06/30/2020, 4:27 PMDarragh
06/30/2020, 4:55 PM