openclaw - 💡(How to fix) Fix [Bug] workflow agent runs causing nested timeout cascading failures [1 participants]

[diagnostic] lane task error: lane=session:agent:main:cron:76c9acdf-1f41-4bbe-b584-f9b956d05728 durationMs=119986 error="FailoverError: LLM request timed out." [diagnostic] lane task error: lane=nested durationMs=119985 error="FailoverError: LLM request timed out."

TL;DR

Implementing a retry mechanism for isolated sessions and containing failures within nested agents can prevent cascading timeouts.

Guidance

• Review the current implementation of nested agent runs to identify where the failure propagation is occurring and consider adding error handling to contain the failure.
• Investigate the possibility of implementing a retry mechanism for isolated sessions that time out, to prevent the outer workflow from failing.
• Analyze the logging evidence to understand the sequence of events leading to the cascading timeout and identify potential points of intervention.
• Consider modifying the workflow configuration to allow for more robust error handling and failure containment.

Example

# Pseudo-code example of a retry mechanism for isolated sessions
def run_isolated_session():
    max_retries =