Qdrant N8N update question

Bug Description

The @n8n/n8n-nodes-langchain.vectorStoreQdrant node (v1.3, retrieve-as-tool mode, attached to an AI Agent) throws "fetch failed" at execution time. The Qdrant cluster is healthy and the credential tests successfully. The same Qdrant endpoint and API key work fine from outside n8n (curl / direct REST), so the collection, key, and network target are all valid. The failure is isolated to this node's fetch path.

To Reproduce

To Reproduce Add a Qdrant Vector Store node in mode "retrieve-as-tool". Attach an OpenAI Embeddings node + a Qdrant credential (Qdrant Cloud). Connect it as a tool to an AI Agent node. Execute the workflow so the agent calls the tool. Node fails with "fetch failed".

Expected behavior

The node should query the Qdrant collection and return the top-K matches to the agent, as it did previously with the same credential and collection.

Debug Info

NodeApiError: fetch failed at normalizeVectorStoreError (.../@n8n/ai-utilities/.../vector-store/createVectorStoreNode/utils.ts:57:8) at ExecuteContext.execute (.../createVectorStoreNode.ts:471:30) ... nodeName: maste-new-collection3 nodeType: @n8n/n8n-nodes-langchain.vectorStoreQdrant nodeVersion: 1.3 n8n version: 2.35.4

Deployment: n8n Cloud Node version: @n8n/n8n-nodes-langchain.vectorStoreQdrant v1.3 Database: (Cloud default — leave as "n8n Cloud") Execution mode: (main / queue — Cloud is queue; put "n8n Cloud") Running via: n8n Cloud Operating system: n8n Cloud (browser: Windows 10) Qdrant: Qdrant Cloud, cluster status Healthy

Operating System

Windows 10

n8n Version

2.38.6

Node.js Version

1.3

Database

SQLite (default)

Execution mode

main (default)

Hosting

n8n cloud