error-recovery

Handle transient API call failures by applying retry with backoff patterns, which automatically retry failed operations using increasing delays and logging to preserve progress and minimize downtime.

Graceful degradation is a recovery pattern that continues delivering value when non-critical workflow steps fail, allowing agents to proceed with partial data while logging clear notes on skipped operations.

Implement fallback alternatives by switching agents to secondary tools or methods automatically when the primary approach fails, ensuring build workflows continue running without manual intervention.

Escalate errors with context by summarizing automated recovery attempts, encountered errors, and recommended next steps, providing humans the exact information needed for manual debugging.

Yes, diagnose-and-fix patterns analyze error output to identify root causes like missing dependencies, then apply targeted fixes to remediate the failure and restore agent coordination automatically.

Retry with backoff targets transient errors only; persistent partial data errors require diagnose-and-fix or graceful degradation patterns to analyze root causes or continue with reduced functionality.