Skip to content

execution_timeout seems to be ignored in EcsRunTaskOperator when using deferable #32580

@tim-x-y-z

Description

@tim-x-y-z

Apache Airflow version

2.6.3

What happened

It seems that with the new aws provider package, when using the deferable keyword in the EcsRunTaskOperator - the execution_timeout is ignored and the task is killed from another timeout, the trigger timeout seems to be timeout=timedelta(seconds=self.waiter_max_attempts * self.waiter_delay + 60).

Also, it seems when the trigger fires that timeout - it seems the task return "success" even though it hasn't finished.
It seems this doesn't kill the task either.

What you think should happen instead

The execution_timeout should be used in the trigger timeout, or at least a warning if that timeout is overriden or is smaller

How to reproduce

Run an EcsRunTaskOperator task with deferable mode, put a large execution_timeout and a small number of waiter_retries. The task should terminates based on the trigger timing out before the execution_timeout is up.

Operating System

linux ubuntu

Versions of Apache Airflow Providers

No response

Deployment

Other Docker-based deployment

Deployment details

No response

Anything else

No response

Are you willing to submit PR?

  • Yes I am willing to submit a PR!

Code of Conduct

Metadata

Metadata

Assignees

No one assigned

    Type

    No type

    Projects

    No projects

    Milestone

    No milestone

    Relationships

    None yet

    Development

    No branches or pull requests

    Issue actions