Support & FAQ

Troubleshooting

Workflow Not Triggering

Symptoms: Workflow is active but never executes.

Checklist:

  1. Confirm the workflow status shows Active (green dot in Workflows list)
  2. Open the trigger node — verify credentials are not expired
  3. For webhooks: send a test request and check the Execution Log for a new entry
  4. For IMAP: verify the mailbox connection in Credentials → Test
  5. For schedules: confirm the cron expression is valid and the timezone is correct

Common cron mistakes:

# Wrong — runs every minute (probably not intended)
* * * * *

# Right — runs at 9:00 AM every weekday
0 9 * * 1-5

Node Returning Empty Data

Symptoms: A node produces {} or [] output unexpectedly.

Steps:

  1. Click the node → Test Node to inspect raw output
  2. Check if the upstream node is actually returning data
  3. Verify field name mappings — field names are case-sensitive
  4. For HTTP Request nodes: check the API response in the Execution Inspector

AI Agent Looping or Not Stopping

Symptoms: Agent runs for many iterations and never produces a final answer.

Fix:

Credential Authentication Failures

Symptoms: 401 Unauthorized or 403 Forbidden errors from external services.

Steps:

  1. Go to Settings → Credentials → [Credential Name] → Test
  2. If OAuth2: click Reconnect to re-authorise
  3. Check if the API key has been rotated on the service's side
  4. Verify the credential has the required scopes/permissions

SAP MCP Connection Issues

Symptoms: SAP tools fail with connection or authentication errors.

Steps:

  1. Verify the SAP MCP Server is running: curl http://your-mcp-host:3001/health
  2. Check the server logs for SAP RFC or OData errors
  3. Confirm the SAP RFC user has the required authorisation objects
  4. For S/4HANA Cloud: verify the OAuth2 token URL and client credentials are correct
  5. Re-test the connection in Settings → MCP Servers → [Server] → Test

High Execution Times

Symptoms: Workflow takes much longer than expected.

Investigate:

  1. Open the Execution Inspector — look for the slowest node
  2. For HTTP nodes: check if the external API is slow (add a timeout)
  3. For AI nodes: model latency is normal (2–10s per call); consider caching results
  4. For loops over large arrays: enable Batch Mode to process N items at a time

Still Need Help?

If you can't resolve the issue from this guide, [contact our support team](/support) with: