flowise-worker failed in queue "flowise-queue-upsertion

[AlterDevi]

Hello All,

I have this message trying to upsert document in document store using local embedding model.

All configured and tested ok but the flowise-worker

"The default method for hashing keys is insecure and will be replaced in a future version,
but hasn't been replaced yet as to not break existing caches. It's recommended that you use
a more secure hashing algorithm to avoid cache poisoning."

See this page for more information:
|
└> https://js.langchain.com/docs/troubleshooting/warnings/insecure-cache-algorithm
2025-10-11 18:04:40 [INFO]: [BaseQueue] Processing job 33 in flowise-queue-upsertion at 2025-10-11T18:04:40.800Z
2025-10-11 18:04:40 [ERROR]: [BaseQueue] Worker job 33 failed in queue "flowise-queue-upsertion":

How can be solved?

Thank you.

[AlterDevi]

Anyone?

How can I correct this: "Insecure Cache Algorithms" from https://js.langchain.com/docs/troubleshooting/warnings/insecure-cache-algorithm

Best regards,

[Ganesh-403]

@AlterDevi
The "insecure cache algorithm" message is most likely not the reason your upsertion job is failing.

That warning comes from LangChain JS and is emitted when the default cache key hashing algorithm is used. It's a warning, not an error, and by itself it should not cause the worker job to fail. The LangChain troubleshooting page also describes it as a security recommendation for cache hashing rather than a runtime failure. :contentReference[oaicite:0]{index=0}

The more important part is this:

[INFO]: Processing job ...
[ERROR]: Worker job 33 failed in queue "flowise-queue-upsertion"

Unfortunately, the actual exception is missing from the log you posted, so it's difficult to determine the root cause.

A few things I'd check:

1. Enable more verbose logging for the worker and capture the full stack trace after:

   Worker job 33 failed in queue "flowise-queue-upsertion"
   

   The exception immediately following that line is usually what identifies the real problem.

2. Verify the embedding model independently.
   Since you're using a local embedding model, confirm it responds correctly outside the worker (e.g. via the "Test" button or direct API call). If the worker cannot reach the embedding endpoint, the upsertion job will fail even though the initial configuration test succeeds.

3. Check the worker configuration.
   Make sure the flowise-worker has access to the same environment variables, model endpoint, and network as the main Flowise server. In distributed deployments it's common for the UI/server to work while the worker cannot reach the embedding service. Flowise's worker architecture processes document upsertion asynchronously through the queue. :contentReference[oaicite:1]{index=1}

Regarding the LangChain warning itself:

"The default method for hashing keys is insecure..."

From what I can tell, I couldn't find any Flowise configuration that lets you suppress or "fix" this warning directly. It appears to originate from the underlying LangChain library, so resolving it would require Flowise (or your custom integration) to adopt the newer cache hashing mechanism once supported. At the moment, I don't see documentation indicating that users need to change anything for normal operation.

Could you share a few more details?

• Which Flowise version are you running?
• Which vector database are you using (FAISS, Chroma, PGVector, Qdrant, etc.)?
• Which local embedding model/provider are you using (Ollama, LM Studio, OpenAI-compatible server, Sentence Transformers, etc.)?
• Can you post the full worker stack trace after Worker job 33 failed?

That stack trace will likely reveal the actual cause of the failed upsertion.

If this solves your problem, feel free to mark it as the accepted answer so others can find it easily.