Hello N8N Community,
I’m facing a persistent issue with a specific workflow on N8N Cloud, and my debugging attempts haven’t been successful. I would greatly appreciate any help or insight.
Workflow Goal (Chat DocGrowX V3):
- Triggered by a Webhook (GET method, receiving
nameandemailas query parameters, e.g.,/webhook/docgrowx-v3?name=X&email=Y). - Use an
Edit Fieldsnode to prepare input. - Use a
Simple Memorynode (withSession IDbased onemail). - Call the OpenAI Assistants API using the
OpenAInode (Message an assistant), passing thename, using theThread IDfrom memory. - Use a second
Edit Fieldsnode (Edit Fields1) to prepare data for logging. - Save the interaction to
Google Sheets(Append Row). - Respond to the initial call using a
Respond to Webhooknode.
The Problem:
When trying to access the Production URL of this workflow’s Webhook (e.g., https://[my-subdomain].app.n8n.cloud/webhook/docgrowx-v3?name=Test&[email protected]), I consistently get the error: {"code":0,"message":"Workflow Webhook Error: Workflow could not be started!"}
The Strange Part:
- A brand new, simple test workflow, on the same N8N Cloud instance, works perfectly. I created a flow with
Webhook(Path:/simple-test, Method:GET, Respond:Immediately) →Set(setting a simple field). Calling its Production URL runs successfully and returns{"message":"Workflow was started"}. This indicates the N8N instance is functional for basic webhooks.
Debugging Steps Already Taken (without success on the Chat DocGrowX workflow):
- Rebuild: Tried recreating the workflow from scratch (V2, V3).
- Extreme Simplification: Reduced the
Chat DocGrowX V3workflow to the absolute minimum:Webhook -> Edit Fields, with all other nodes deactivated. The “could not be started!” error persisted. - Webhook
RespondConfiguration:
- Tried
Respond: Immediately(with subsequent nodes deactivated). The “could not be started!” error continued. - Tried
Respond: Using 'Respond to Webhook' Node'. This correctly generated the “No ‘Respond to Webhook’ node found…” error when the response node was deactivated, but still resulted in the “could not be started!” error when the response node was active (suggesting the startup issue occurred earlier).
Simple MemoryConfig: Found and fixed an error in theSession IDconfiguration of theSimple Memorynode, but this did not resolve the startup issue.- Activation/Saving: Ensured multiple times that the workflow was saved and active (green “Active” toggle), including deactivating/reactivating after each change.
- Intermediate Node Config: Reviewed configurations of
Edit FieldsandOpenAInodes to ensure no obvious errors that might prevent N8N’s initial validation (although the failure occurs before they execute). - Invalid JSON Error: During debugging, I also encountered a persistent invalid JSON error (
{"{ \"mensagem\": $json...) that seemed to originate fromEdit Fields1orRespond to Webhook, but this error only appeared when the workflow managed to start, which is no longer happening. The primary blocker now is the failure to start.
Current Conclusion:
There seems to be a specific issue with this workflow (Chat DocGrowX V3) or perhaps the path /docgrowx-v3 that prevents N8N from initiating it correctly, even in its simplest configuration, despite other webhooks working on the same instance. I suspect some internal workflow corruption or a specific hidden issue.
Has anyone encountered such a persistent and specific “Workflow could not be started!” error? Any suggestions on where else to look or what else to try?
Thanks! (Optional: I can provide the workflow JSON if helpful.)