@Glime I am running a

map_task

within a dynamic workflow. I have set the

timeout

using

flytekit.TaskMetadata

and passing it as

metadata

to

map_task

. My expectation was that it would apply to each of the tasks executed by

map_task

, but I am getting timeout errors when the whole

map_task

runtime breaches the same timeout. Was my understanding not correct?

Your understanding was partially correct. In Flyte, when you set a

timeout

using

flytekit.TaskMetadata

and pass it to a

map_task

, the expectation is indeed that it applies to the individual tasks executed by the

map_task

. However, there seems to be a known issue where the

timeout

is incorrectly applied to the entire

map_task

node, rather than to each individual task within it. This behavior is not expected and has been recognized as a bug in the Flyte system. This issue has been discussed in the Flyte community, where it was noted that the

timeout

applies to the entire

map_task

node when it should ideally apply to each task individually. This can cause the

map_task

to fail if the total execution time of all tasks sequentially exceeds the specified timeout, even if each individual task does not exceed the timeout limit on its own. To address your issue, it would be beneficial to follow the updates on this bug and check for any patches or updates that might resolve this incorrect timeout behavior. In the meantime, you might need to adjust your workflow design or timeout settings to accommodate this behavior. --- This is an AI-generated response, take its recommendations with caution. Please help us improve the model accuracy by giving this response a thumbs up/thumbs down reaction. Sources https://github.com/flyteorg/flyte/issues/3875/ 👍 button 👎 button