# Databricks MCP for AI Agents AI Agent Connect

> Databricks MCP lets you manage your lakehouse and data intelligence platform directly through your AI agent. You can monitor compute clusters, track job executions, audit SQL warehouses, and explore the Unity Catalog without switching tabs. It gives your agent the eyes and hands it needs to interact with your Databricks workspace.

## Overview
- **Category:** industry-titans
- **Price:** Free
- **Endpoint:** https://edge.vinkius.com/vk_preview_aLUngwuWdCUusUCfiR78zXA9b27HYNx59jOLRy9C/ai-agent-connect
- **Tags:** lakehouse, data-engineering, cluster-management, sql-warehousing, data-governance, big-data

## Description

Managing a complex data lakehouse usually means jumping between different dashboards, checking cluster statuses, and digging through job logs to find out why a pipeline failed. It's a lot of context switching that breaks your flow. This Connector changes that by giving your AI agent direct access to your Databricks workspace. Instead of manually hunting for information, you can just ask your agent to pull up the status of your active clusters or find the specific schema where your data lives. It handles the heavy lifting of navigating the Unity Catalog and checking warehouse boundaries so you can stay focused on the actual data work. Because Vinkius hosts this Connector in a central catalog, you can connect it to any of your favorite AI clients and get these capabilities across your entire workflow instantly. It turns your agent into a capable data platform operator that can audit permissions, list workflows, and troubleshoot job failures on demand.

## Tools

### list_warehouses
Enumerates all configured SQL Serverless warehouses. This lets you see your available compute resources.

### list_clusters
Lists all compute nodes in your workspace. Use this to see what's running and what's not.

### get_cluster
Fetches specific details for a single cluster. This helps you audit health and capacity limits.

### list_jobs
Lists all configured workflows and jobs. Use this to see your entire pipeline lineup.

### list_job_runs
Lists the history of job executions. This is how you find out why a specific run failed.

### list_catalogs
Lists all root catalogs in the Unity Catalog. This is the first step to finding your data.

### list_schemas
Lists all schemas within a specific catalog. This helps you see your database structure quickly.

### get_me
Retrieves the profile of the currently authenticated user. Use this to check your active permissions.

## Prompt Examples

**Prompt:** 
```
What's the status of our compute clusters?
```

**Response:** 
```
### Cluster Status Summary

| Cluster Name | Status | Capacity |
| :--- | :--- | :--- |
| **Standard-ETL** | ✅ Running | 80% |
| **ML-Training** | ❌ Terminated | 0% |
| **Shared-Dev** | ✅ Running | 45% |

Would you like me to pull the full hardware specs for any of these?
```

**Prompt:** 
```
Show me the last 3 runs for the 'Daily_Sync' job.
```

**Response:** 
```
### Job Run History: Daily_Sync

*   **Run ID 987**: ✅ Success (6:00 AM)
*   **Run ID 985**: ❌ Failed (5:30 AM) - *Reason: Cluster Timeout*
*   **Run ID 983**: ✅ Success (4:00 AM)

I can pull the detailed error logs for Run 985 if you'd like.
```

**Prompt:** 
```
List the catalogs in our Unity Catalog.
```

**Response:** 
```
I found the following root catalogs in your Unity Catalog:

*   **main**
*   **development**
*   **samples**
*   **system**

Which one should I explore for specific databases or schemas?
```

## Capabilities

### Monitor compute cluster health
See the status and capacity of all your nodes at once.

### Track job execution statuses
Check if your data pipelines finished successfully or hit a snag.

### Explore Unity Catalog structures
Quickly find where specific databases and schemas are located.

### Audit SQL warehouse limits
View the operational boundaries of your serverless warehouses.

### Verify user permissions
Check your own profile or service principal access levels.

### Retrieve job run logs
Pull chronological history to pinpoint exactly where a workflow failed.

## Use Cases

### Troubleshooting a failed ETL
A data engineer asks the agent to show the last 5 runs for a specific job and then pull the error logs.

### Auditing cluster capacity
A platform lead asks the agent to list all clusters and show the details for the ML training nodes.

### Data discovery
An analyst asks the agent to list all catalogs and then find the schemas for the marketing database.

### Permission checks
A developer uses the agent to check if their service principal has the right permissions to access a specific catalog.

## Benefits

- Stop manual dashboard hunting by using `list_clusters` to see your compute status instantly.
- Find failed data pipelines faster by using `list_job_runs` to pull specific error logs.
- Navigate the Unity Catalog easily with `list_catalogs` and `list_schemas` to locate data.
- Audit your SQL resources quickly with `list_warehouses` to check serverless boundaries.
- Verify your own access levels on the fly using `get_me` to check permissions.

## How It Works

The bottom line is you get a direct line of communication between your AI agent and your Databricks workspace.

1. Subscribe to the Databricks MCP on Vinkius.
2. Provide your Databricks Host URL and Personal Access Token (PAT).
3. Ask your AI agent to list clusters, check jobs, or explore catalogs.

## Frequently Asked Questions

**Can the Databricks MCP help me find failed jobs?**
Yes, it can list all your configured jobs and pull the execution history for specific runs. You can then ask your agent to show you the logs for any failed run to see exactly what went wrong.

**Does the Databricks MCP work with the Unity Catalog?**
It does. You can use it to list all root catalogs and drill down into specific schemas and databases to find your data without leaving your agent.

**How does the Databricks MCP help with cluster management?**
You can use it to see every compute cluster in your workspace at once. It also lets you pull detailed health and capacity information for any specific cluster.

**Can I use the Databricks MCP to check my permissions?**
Yes, the Connector can fetch your current profile information. This is useful for verifying your active permissions or checking a service principal's access levels.

**Does the Databricks MCP support SQL warehouses?**
Yes, it can enumerate all your configured SQL Serverless warehouses. This helps you keep track of your available SQL compute resources.

**Can my agent check the status of a specific Databricks job run?**
Yes. Provide the 'job_id' to the 'list_job_runs' tool. The agent will retrieve the chronological history of executions, allowing you to identify successful completions or precise points of failure in your workflows.

**How do I explore schemas within a specific Unity Catalog via chat?**
Use the 'list_schemas' tool and provide the catalog name. Your agent will pull the detailed databases and schemas registered inside that Unity Catalog, giving you immediate visibility into your data hierarchy.

**Can I monitor the health of my Databricks clusters through the agent?**
Absolutely. The 'list_clusters' and 'get_cluster' tools allow your agent to retrieve detailed node information and operational statuses, helping you audit cluster health and capacity across your workspace.