Error Messages

This reference explains error messages you may encounter and how to resolve them.

Validation Errors

These errors occur when saving or releasing workflows.

"Node is not connected"

Meaning: A node has no incoming or outgoing connections.

Resolution:

1. Find the disconnected node
2. Connect it to other nodes
3. Or remove it if not needed

"Missing required field: [field]"

Meaning: A required configuration field is empty.

Resolution:

1. Open the node configuration
2. Fill in the required field
3. Save the configuration

"Invalid CEL expression"

Meaning: The CEL expression has syntax errors.

Resolution:

1. Check the expression syntax
2. Verify property names match your data
3. Check parentheses and operators
4. See CEL Reference

"Invalid Liquid template"

Meaning: The Liquid template has syntax errors.

Resolution:

1. Check for unclosed tags ({% %} or {{ }})
2. Verify filter syntax
3. Check for typos in variable names
4. See Liquid Reference

"Trigger not configured"

Meaning: The workflow has no trigger.

Resolution:

1. Add an Event Trigger or Scheduled Trigger
2. Configure the trigger type
3. Connect to downstream nodes

"Circular reference detected"

Meaning: The workflow graph contains a loop.

Resolution:

1. Review node connections
2. Remove the edge creating the cycle
3. Restructure workflow if needed

Execution Errors

These errors occur during workflow execution.

Timeout Errors

"Execution timed out after [X]ms"

Meaning: Node took too long to complete.

Default timeouts:

• Standard nodes: 30 seconds
• AI nodes: 120 seconds

Resolution:

1. Reduce input size
2. Simplify the operation
3. Check external service response times
4. Consider splitting into multiple nodes

"AI request timed out"

Meaning: AI model didn't respond in time.

Resolution:

1. Use transcript summary instead of full transcript
2. Reduce prompt complexity
3. Try at a different time
4. Use lower model tier

Data Errors

"Property '[name]' not found"

Meaning: Trying to access a property that doesn't exist.

Resolution:

1. Check the property name spelling
2. Verify upstream node produces this property
3. Add null check before accessing
4. Check execution log for actual data structure

"Cannot read property of null"

Meaning: Trying to access property on null value.

Resolution:

1. Add null check in condition
2. Use If node to handle null case
3. Verify upstream data exists

"Index out of bounds"

Meaning: Array index doesn't exist.

Resolution:

1. Check array size before accessing
2. Use size() > 0 check
3. Handle empty array case

"Type mismatch: expected [type] got [type]"

Meaning: Value is wrong type for operation.

Resolution:

1. Use type conversion functions
2. Check data sources
3. Verify return type configuration

Integration Errors

"Authentication failed" / "401 Unauthorized"

Meaning: Integration credentials are invalid.

Resolution:

1. Go to Settings → Integrations
2. Disconnect the integration
3. Reconnect with fresh credentials

"Permission denied" / "403 Forbidden"

Meaning: Account doesn't have required permissions.

Resolution:

1. Check integration permissions
2. Verify account access level
3. Reconnect with appropriate scopes

"Resource not found" / "404 Not Found"

Meaning: The requested object doesn't exist.

Resolution:

1. Verify the ID is correct
2. Check the object wasn't deleted
3. Confirm object type matches

"Rate limited" / "429 Too Many Requests"

Meaning: Too many API calls in short time.

Resolution:

1. Add Wait nodes between calls
2. Reduce request frequency
3. Implement backoff strategy

"Service unavailable" / "503"

Meaning: External service is down.

Resolution:

1. Wait and retry later
2. Check service status page
3. Message will retry automatically

AI Errors

"Failed to parse AI output"

Meaning: AI response doesn't match expected format.

Resolution:

1. Add clearer format instructions to prompt
2. Include output examples
3. Use simpler return type
4. Add "Return only the [type], nothing else"

"Invalid JSON in AI response"

Meaning: Expected JSON but got malformed response.

Resolution:

1. Add explicit JSON format instructions
2. Include JSON example in prompt
3. Use string return type and parse separately

"AI output exceeds maximum length"

Meaning: Response too long.

Resolution:

1. Add length constraints to prompt
2. Request shorter output
3. Split into multiple requests

Retry and DLQ Errors

"Max retries exceeded"

Meaning: All retry attempts failed.

Resolution:

1. Check the original error
2. Fix root cause
3. Consider retry from DLQ

"Message moved to DLQ"

Meaning: Message couldn't be processed.

Resolution:

1. Review DLQ for details
2. Fix the underlying issue
3. Retry if appropriate

System Errors

"Internal server error"

Meaning: Unexpected platform error.

Resolution:

1. Wait and retry
2. Check if reproducible
3. Contact support if persistent

"Workflow version not found"

Meaning: Referenced version doesn't exist.

Resolution:

1. Check version exists
2. Ensure it's not deleted
3. Reference correct version

"Execution not found"

Meaning: Execution ID doesn't exist.

Resolution:

1. Verify execution ID
2. Check it wasn't purged
3. Use correct organization

Error Code Reference

Reading Error Messages

Error Structure

Errors typically include:

[Error Code] [Category]: [Message]

Details:
- Node: [node name]
- Input: [relevant input]
- Context: [additional info]

What to Look For

1. Error code - Identifies error type
2. Message - Human-readable description
3. Node - Where the error occurred
4. Details - Additional context

Example Error Analysis

EXE_002 Execution: Property 'callRecording' not found

Details:
- Node: AI Prompt - Summarize
- Input: { meeting: {...} }
- Context: Check if meeting has a recording

Analysis:

1. Error occurred in "AI Prompt - Summarize" node
2. The input has meeting but no callRecording
3. The meeting likely doesn't have a recording

Solution: Add an If node to check json.callRecording != null before the AI node.

Related

• Common Issues
• FAQ
• Error Handling Guide