- 158 Views
- 2 replies
- 1 kudos
How to deploy an Agent
Hi everyone, I’m currently developing several Databricks Agents using a custom Claude skill to integrate Databricks Asset Bundles (DABs), Mosaic AI frameworks, and LangGraph. My current workflow is the following: I develop the agents locally.For depl...
- 158 Views
- 2 replies
- 1 kudos
- 1 kudos
Hey, your approach is a common pattern. But there has been a recent shift where we now recommend creating the agents as databricks apps instead. This allows you to not do pipeline execution steps. Here are a couple of articles you may find helpful. h...
- 1 kudos
- 6054 Views
- 3 replies
- 1 kudos
Behavior of Vector Index Sync with Delta Tables When Using OVERWRITE vs MERGE in Databricks
I'm working with vector search in Databricks using vector index sync with Delta tables, and I'm a bit unclear on how updates to the source table affect the vector index, specifically when using different write operations.If I overwrite the source Del...
- 6054 Views
- 3 replies
- 1 kudos
- 1 kudos
From community experience, vector index sync behavior depends heavily on how the Delta table is updated. With OVERWRITE, the table is effectively replaced, so the vector index typically treats this as a full refresh. Existing embeddings are dropped a...
- 1 kudos
- 472 Views
- 1 replies
- 0 kudos
Genie - Value dictionary
When I add the tables to the genie space, it automatically turns on the value dictionary for first 120 string fields, Is there a way to disable them by default and add only to the needed fields later?I am working in curating responses for around 15 t...
- 472 Views
- 1 replies
- 0 kudos
- 0 kudos
"Value sampling is enabled by default for all Genie spaces." The only solution that I see is to build json and use as a serialized_space template https://docs.databricks.com/api/azure/workspace/genie/getspace
- 0 kudos
- 565 Views
- 3 replies
- 4 kudos
Multi‑Agent Supervisor: url_citation (source links) not shown in Playground — why and how to enable?
I’m seeing a difference in citation behavior between a single Knowledge Assistant (KA) agent and the Multi‑Agent Supervisor setup.What I testedIn Agent Bricks → Knowledge Assistant, I created an agent that returns citations with links (e.g., url_cita...
- 565 Views
- 3 replies
- 4 kudos
- 552 Views
- 1 replies
- 1 kudos
Getting error while trying to use external MCP
I have created a mcp connection from marketplace :Conn Name : mcp_brightdata_web_surfermcp product url : https://brightdata.com/I have attached the mcp through playground . But while submitting the prompt I am getting the following error .Error : The...
- 552 Views
- 1 replies
- 1 kudos
- 1 kudos
Based on the error, which is a "502 Bad Gateway" error indicates that a server sitting as a gateway or proxy received an invalid response from an upstream server. Here, Databricks is trying to communicate with the external MCP server associated with...
- 1 kudos
- 3745 Views
- 1 replies
- 0 kudos
Libraries in requirements file not installed when streamlit app is deployed
I deployed a streamlit app on Databricks cluster but the libraries in the requirments.txt file were not installed when the app is deployed. I'm getting this error when a function that uses OpenAI is called."ImportError: Could not import openai python...
- 3745 Views
- 1 replies
- 0 kudos
- 0 kudos
Hi @data_scientist3 , Would you share how you deployed the Streamlit app (Databricks Apps vs running Streamlit from a notebook/job cluster) and your cluster’s Databricks Runtime version? That’ll help me tailor the exact steps to your setup. Option 1:...
- 0 kudos
- 1036 Views
- 3 replies
- 0 kudos
Resolved! Testing out Agentic Capabilities
So I am creating a POV on Databricks' agentic capabilities and wanted to showcase its abilities through a simple change pipeline.A user asks for changes in a specific table in a schema -> based on metadata info from our lake table info is received ->...
- 1036 Views
- 3 replies
- 0 kudos
- 0 kudos
Yes, it’s viable — Databricks’ in-house agents can handle that workflow if you define each stage as callable Python tools. The key is robust function design and proper metadata access. However, you might find integrating Unity Catalog and Git integra...
- 0 kudos
- 1763 Views
- 1 replies
- 1 kudos
Permission Denied for Genie Auto-Generated Service Principal on SQL Endpoint in Playground
Hi community,Use Genie in multi-agent systems | Databricks DocumentationI’ve developed a multi-agent Genie in Databricks and integrated it with vector indexes. The setup works fine during model logging and prediction. The system successfully register...
- 1763 Views
- 1 replies
- 1 kudos
- 1 kudos
Hi @Karthik_Karanm - Can you ensure to add the Genie in the resources as mentioned in the TODO of the cell. To enable automatic authentication, specify the dependent Databricks resources when calling mlflow.pyfunc.log_model(). TODO: If your Unity Ca...
- 1 kudos
- 3986 Views
- 1 replies
- 0 kudos
Multi-Vector Embedding Support
Does Databricks Vector Search support Multi-Vector Embedding?
- 3986 Views
- 1 replies
- 0 kudos
- 0 kudos
I believe we can only have one embedding source column mapped to one embedding column.
- 0 kudos
- 1658 Views
- 1 replies
- 0 kudos
Vector Quantization
Does Databricks Vector Search support vector quantization?
- 1658 Views
- 1 replies
- 0 kudos
- 0 kudos
Hi SaiVihAIIn my opinion, Databricks utilizes Mosaic Vector Search, which is largely aided by the Hierarchial Navigable Small World (HNSW) algorithm for neighbor search and hybrid keyword similarity search.Ref Mosaic AI Vector Search | Databricks D...
- 0 kudos
- 1712 Views
- 3 replies
- 0 kudos
Should elaborate and complex LLM apps be deployed as MLFlow serving endpoints?
In a project we are building increasingly complex LLM-based Apps (RAG, multi-agent workflows, langgraph, unstructured ingestion etc), and we are having doubts if these apps should be deployed as MLFlow-based endpoints. I would like your feedback on i...
- 1712 Views
- 3 replies
- 0 kudos
- 0 kudos
Actually, maybe root folder was imprecise. The point is that it gets file system access. It becomes a regular Workspace user, with too much access. If, however, you want to give it specific accesses beyond that, you could give it access to specific v...
- 0 kudos