I’m getting stuck trying to set up a basic N8N workflow with Qdrant vector database. When I run my workflow, it throws this error:
Problem in node 'Qdrant Vector Store'
400 Bad Request: Wrong input: Not existing vector name error:
My workflow is pretty straightforward - it uses OpenAI embeddings, loads documents from a Git repository, splits the text, and tries to store everything in Qdrant. But something is wrong with how I configured the vector store.
I created a collection in Qdrant dashboard but maybe I’m missing some settings that make it compatible with N8N. The documentation doesn’t really explain the collection setup process clearly.
Has anyone successfully connected N8N with Qdrant before? What collection settings should I use in the Qdrant dashboard to make this work properly?
Just hit this yesterday. The problem’s probably deeper than collection naming - check your Qdrant collection’s vector config. When you made the collection in the dashboard, did you set vector size to 1536? That’s critical for OpenAI embeddings. Also make sure your N8N workflow’s actually connecting to the right Qdrant instance. I screwed up by running multiple Qdrant containers and N8N was hitting the wrong one. Test the connection manually with a curl command to your Qdrant endpoint before running the full workflow. Sometimes the collection looks created but isn’t properly initialized with the vector schema N8N needs.
The Problem:
You’re encountering issues connecting n8n to your Qdrant vector database, resulting in a “400 Bad Request: Wrong input: Not existing vector name” error. This indicates a problem with how your n8n workflow interacts with your Qdrant collection, specifically concerning the vector field name. The issue is not solely related to collection setup, but rather involves the integration between n8n and Qdrant.
Understanding the “Why” (The Root Cause):
n8n’s built-in nodes for interacting with vector databases like Qdrant can be inflexible and may not handle all edge cases or nuanced configurations seamlessly. The “Not existing vector name” error highlights a mismatch between the vector field name expected by Qdrant and the name specified in your n8n configuration. This discrepancy arises from either incorrect naming in your Qdrant collection setup, inaccurate mapping in the n8n node configuration, or problems with the overall communication between n8n and the Qdrant instance.
Step-by-Step Guide:
-
Migrate to a More Flexible Solution (Recommended): The most effective and efficient solution is to bypass n8n’s limitations by directly interacting with the Qdrant API using a more flexible tool that offers better control over data transfer and error handling. This allows you to create requests tailored exactly to Qdrant’s requirements, ensuring compatibility with your specific collection setup and preventing naming mismatches.
-
Verify Qdrant Collection Setup (If Not Migrating): Access your Qdrant dashboard and meticulously check the name of the vector field in the collection you’re targeting. Ensure this name is accurate and precisely matches the name you’ve specified in your n8n Qdrant node. Note that Qdrant is case-sensitive, even though the n8n node interface might not always be.
-
Double-Check n8n Node Configuration (If Not Migrating): In your n8n workflow, carefully review the configuration of the Qdrant node. Specifically, confirm that the vector field name used in the n8n node precisely matches the name you verified in your Qdrant dashboard. Even minor discrepancies (e.g., extra spaces, capitalization differences) will cause the error.
-
Test Connection Manually (If Not Migrating): Before running the complete workflow, test the connection to your Qdrant instance directly using a tool like curl. This helps isolate whether the problem is with your n8n configuration or with Qdrant itself. A successful curl request confirms that n8n can reach Qdrant; if it fails, troubleshoot network connectivity and Qdrant instance availability.
-
Inspect n8n Logs (If Not Migrating): If problems persist, examine the n8n execution logs thoroughly. These logs provide detailed information about the interaction between n8n and Qdrant, including any error messages that might pinpoint the exact cause of the mismatch.
Common Pitfalls & What to Check Next:
-
Vector Dimensions: Ensure the dimensions of your vectors (e.g., 1536 for OpenAI’s Ada embeddings) in your Qdrant collection match the output of your embedding generation process in your n8n workflow. A mismatch in vector size will lead to errors.
-
Distance Metric: Verify that the distance metric you have selected in Qdrant (e.g., cosine similarity) is suitable for your use case and aligns with your embedding model.
-
Collection Initialization: Confirm that your Qdrant collection is correctly initialized and contains the expected vector field. A simple curl request to Qdrant’s points endpoint with a test vector will demonstrate successful communication and collection existence.
-
n8n Version: Check for potential issues caused by older n8n versions. Updating to the latest version often resolves bugs and improves compatibility.
Still running into issues? Share your (sanitized) config files, the exact command you ran, and any other relevant details. The community is here to help!
This error usually happens when your collection setup doesn’t match what N8N expects. Go check your collection’s vector field name in the Qdrant dashboard - it needs to match exactly what you’ve got in your N8N node config. I ran into this same problem and found out Qdrant was looking for a completely different vector field name than what I’d set up. Also double-check that your collection actually exists and N8N can reach it through the URL/port you’re using. Sometimes collection creation fails even when it shows up in the dashboard. Try making a simple test collection with basic vector settings first - that’ll tell you if it’s a naming problem or if there’s something wrong with the connection.
had this exact issue last week! problem was my n8n qdrant node settings - wrong collection name. make sure your collection name in n8n matches exactly what’s in your qdrant dashboard (case sensitive). also check if qdrant’s running on the right port. mine was on 6333 but n8n was trying 6334.
I encountered a similar problem when integrating N8N with Qdrant. The root of the issue was in the configuration of the collection’s vector settings. It’s essential to ensure that your vector dimensions align with those used by OpenAI’s embeddings, which is 1536 for text-embedding-ada-002. Additionally, make sure to set the distance metric to cosine similarity instead of euclidean. If your vector field is labeled ‘vector’, you should verify that N8N is referencing that name correctly. I resolved my issue by recreating the collection with the appropriate vector settings.
This topic was automatically closed 24 hours after the last reply. New replies are no longer allowed.