Workflow Debugging and Troubleshooting

The platform provides comprehensive error handling for workflow execution:

• Automatic retry: Failed nodes can automatically retry based on configuration
• Error routing: Route execution to different nodes when errors occur
• Error skipping: Skip failed nodes and continue execution
• Error logging: Detailed error logs with timestamps, node IDs, and error types
• Execution status: Track node status (success, failed, retry, skip, route_to)

Each node can be configured with error handling settings. You can set retry attempts, retry delays, error routes, and whether to skip on error.

Configure retry logic for unreliable operations:

• Max retries: Maximum number of retry attempts (default 3)
• Retry delay: Delay between retries (exponential backoff supported)
• Retry conditions: Which errors should trigger retries
• Retry status: Execution status shows retry attempts and delays

Retry logic is configured per node. The system automatically retries failed operations with exponential backoff, helping recover from temporary failures like network issues or API rate limits.

Access detailed execution logs to debug issues:

• Execution history: View all past workflow executions
• Node logs: See logs for each node in the workflow
• Error details: Detailed error messages with stack traces
• Timestamps: When each node started and finished
• Execution time: How long each node took to execute
• Input/output data: See what data each node received and produced

Execution logs are available in the execution monitor and can be exported for offline analysis. Logs include all the information you need to debug workflow issues.

The system validates connections as you build workflows:

• Type checking: Validates that data types match between connected nodes
• Connection errors: Shows error messages when connections are invalid
• Visual indicators: Valid connections appear in blue, invalid in red
• Auto-validation: Connections are validated automatically as you make them

Connection validation helps catch errors early, before running workflows. The system checks data types, required fields, and connection compatibility.

Best practices for debugging workflows:

• Test individual nodes: Test nodes one at a time before connecting them
• Check logs: Review execution logs to see what each node did
• Verify inputs: Ensure nodes receive the expected input data
• Check outputs: Verify that nodes produce the expected output
• Use breakpoints: Add condition nodes to pause execution and inspect data
• Simplify workflows: Break complex workflows into smaller, testable parts
• Test with sample data: Use known inputs to verify node behavior

The workflow designer shows real-time cost estimates and connection validation, helping you catch issues before running workflows.