Skip to content
Closed
Changes from all commits
Commits
File filter

Filter by extension

Filter by extension

Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
4 changes: 2 additions & 2 deletions temporalio/contrib/openai_agents/_model_parameters.py
Original file line number Diff line number Diff line change
Expand Up @@ -20,13 +20,13 @@ class ModelActivityParameters:
task_queue: Optional[str] = None
"""Specific task queue to use for model activities."""

schedule_to_close_timeout: Optional[timedelta] = timedelta(seconds=60)
schedule_to_close_timeout: Optional[timedelta] = None
"""Maximum time from scheduling to completion."""

schedule_to_start_timeout: Optional[timedelta] = None
"""Maximum time from scheduling to starting."""

start_to_close_timeout: Optional[timedelta] = None
start_to_close_timeout: Optional[timedelta] = timedelta(seconds=60)
Copy link
Member

@cretz cretz Jul 15, 2025

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

IMO we should not do this until we are also properly differentiating retryable vs non-retryable errors on the model activity. If someone has an auth failure, it will now retry forever. Can we except-and-re-raise model activity failures as application error properly accounting for what is retryable and what's not (basically 4xx HTTP is non-retryable, 5xx is retryable, or some variant of that logic)?

"""Maximum time for the activity to complete."""

heartbeat_timeout: Optional[timedelta] = None
Expand Down
Loading