Troubleshooting
Agent Actions provides actionable error messages and multiple debugging options.
Logging Levels
| Environment Variable | Level | Shows |
|---|---|---|
| (default) | CRITICAL | Only critical system errors |
AGENT_ACTIONS_LOG_LEVEL=INFO | INFO | Progress and status updates |
AGENT_ACTIONS_LOG_LEVEL=DEBUG | DEBUG | Full exception chains, context, and tracebacks |
Debug Mode
Set the AGENT_ACTIONS_LOG_LEVEL environment variable to see full error context:
AGENT_ACTIONS_LOG_LEVEL=DEBUG agac run -a my_workflow
Debug mode reveals:
- Full exception chains showing where errors originated
- Context information (file paths, action names, operations)
- Complete Python tracebacks
tip
Redirect debug output to a file for complex issues: AGENT_ACTIONS_LOG_LEVEL=DEBUG agac run -a my_workflow > debug.log 2>&1
Environment Variables
| Variable | Description |
|---|---|
ANTHROPIC_API_KEY | Anthropic API key |
OPENAI_API_KEY | OpenAI API key |
GEMINI_API_KEY | Google Gemini API key |
GROQ_API_KEY | Groq API key |
MISTRAL_API_KEY | Mistral API key |
COHERE_API_KEY | Cohere API key |
Exit Codes
| Code | Meaning |
|---|---|
0 | Success |
1 | General error |
2 | Usage error (invalid arguments) |
130 | Interrupted by user (Ctrl+C) |
Debugging Process
- Start with
AGENT_ACTIONS_LOG_LEVEL=INFOto see execution flow - Use
AGENT_ACTIONS_LOG_LEVEL=DEBUGto investigate specific errors - Check exception chains for root causes
- Review context information (file paths, action names, operations)
See Also
- run Command - Execute workflows
- schema Command - Validate field references before execution