-
Notifications
You must be signed in to change notification settings - Fork 40
Conversation
I think it'll be important for us to keep in mind that this'll make it easier for users to accidentally supply values to a single container that are larger than the values supplied to the task, especially when they leave the task-level resource settings empty and then request resources for their container that are larger than the default. I'm ok with letting AWS raise the error for us in cases where users specify many of their own containers and get the math wrong, but the case above seems like an easy pitfall for those not familiar with ECS. @jeanluciano Do you think we should check if the Prefect container's resources are larger than the task's resources and reject the run with a nice error? Or maybe add a note on the new settings that they can't exceed the task level resources. Like: |
Sepererates the way we override task level cpu and memory from the prefect container. If no container cpu and/or memory is passed, it uses task level.
Example
Checklist
pre-commit
checks.pre-commit install && pre-commit run --all
locally for formatting and linting.mkdocs serve
view documentation locally.