feat(config): make per-agent execution timeout configurable - #150
Merged
Conversation
The SmolAgentFactory execution timeout was hardcoded to 3600s, which caused workflows that legitimately needed longer to be killed mid-run (Agent '<name>' execution timed out after 3600 seconds), yielding a 0.0 reward with no captured state. Add config.agent_execution_timeout (default 3600s), serialize it via jsonify/from_json/__str__, inject it into the generated workflow header as AGENT_EXECUTION_TIMEOUT, and have SmolAgentFactory read that global (falling back to 3600s when run standalone). Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
Collaborator
|
This is a good idea. Keep in mind that in real world research, any task can take up to days to complete. 1h timeout is definitely too small. |
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
3 participants
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
What
Makes the SmolAgentFactory per-agent execution timeout configurable via
config.agent_execution_timeout(default 3600s / 1h) instead of a hardcoded3600.Why
When a workflow legitimately needs longer than 1h, the agent was killed mid-run (
Agent '<name>' execution timed out after 3600 seconds), producing a 0.0 reward with no captured state — even though the pipeline steps were succeeding. This was observed killing multiple runs back-to-back.How
config.py: newagent_execution_timeoutfield, serialized viajsonify/from_json/__str__.sources/core/workflow_factory.py: injects it into the generated workflow header asAGENT_EXECUTION_TIMEOUT.sources/modules/smolagent_factory.py: reads that global (falls back to 3600s when run standalone).Override per run by adding
"agent_execution_timeout": <seconds>to the--configJSON. Default behavior is unchanged.🤖 Generated with Claude Code