Dbx MCP Server (dbx-mcp-server-9e27137c)

MCP开发工具

提供方 ms-mfg-community

开发工具 stdio · 本地进程 Token 社区 / 第三方 VS Code
查看接入方式

Databricks Error Logs MCP Server 是一款用于在开发环境中搜索和分析 Databricks 中存储的应用程序错误日志的服务器,支持多种运行模式,包括本地开发、本地HTTP服务和Azure托管。

一句话选型结论

按资料字段整理

Dbx MCP Server (dbx-mcp-server-9e27137c) 属于开发工具类 MCP 服务器,stdio · 本地进程,Token,可直接接入 VS Code,有可直接执行的安装命令。接入前先确认权限范围;涉及写入、删除或付款的操作,建议先在测试环境验证。

分类开发工具 传输方式stdio · 本地进程 鉴权Token 部署方式本地部署
GitHub 星标 0 仓库热度,无公开仓库时显示未标注
工具数 7 暴露给模型的工具数量
资源 / 提示词 0 / 0 资源与提示词模板数量
维护方 社区 / 第三方 核验等级 仅资料核对

工具与能力

开发工具
数量来自来源目录的标注,具体工具清单以官方文档为准
工具数量7暴露给模型可调用的工具个数
资源数量0可读取的资源条目
提示词数量0内置提示词模板数量

能力标签

支持客户端

适用场景

错误日志分析Python本地部署Databricks集成开发工具GitHubCopilot扩展日志管理

安装与配置

stdio · 本地进程 Token
客户端在本机拉起子进程,数据不出本机
uv run python -m databricks_error_logs_mcp

这条命令只是入口,具体启动参数请按官方文档补齐,不要直接照搬到别的服务器。

  1. 1
    准备运行环境

    stdio 服务在本地起进程,需要先装好 Python,再按安装命令拉取依赖。

  2. 2
    安装服务

    复制上面的安装命令执行;命令里带占位符的位置,换成自己的密钥或路径。

  3. 3
    写入客户端配置

    在客户端的 MCP 配置里填好启动命令和参数,保存后完全退出客户端再打开,让新配置生效。

  4. 4
    验证是否可用

    连接成功后客户端会列出约 7 个工具,随便调一个只读工具确认能返回结果。

典型使用场景

3 类
按分类与标签整理的常见用法,具体以官方文档为准
代码与仓库操作

在对话里直接查代码、提 issue、走 PR,省掉来回切工具。

本地命令与脚本

把构建、测试、格式化这类重复命令交给模型执行。

接口联调与测试

让模型调你的接口、跑一遍用例,快速定位问题。

选择建议

选之前先看这几条
按条目字段给出的判断,不替你决定用不用

值得选的地方

  • 资料标注 7 个工具,职责相对集中,接入后不容易误操作。
  • 有现成的安装命令,不用自己找入口,照着跑就能起服务。
  • 已适配 VS Code,换客户端不用重新研究配置。

先想清楚的地方

  • stdio 方式要求本地装好运行时(如 Node.js),换机器要重新装一遍。
  • 使用前要准备好Token对应的凭据,并按最小权限配置。
  • 版本更新与行为细节以官方仓库为准。

更适合

  • 需要在开发工具场景里让模型直接调用外部能力的人。
  • 已经在用 VS Code 的用户,接入路径最短。
  • 想先快速试一下、能接受在本地或自有环境跑服务的人。

不太适合

  • 完全不想在本地装环境的人。
  • 不方便管理密钥或第三方授权的人。
  • 需要厂商级 SLA 和技术支持的生产环境。

同类对比

6 个同类
同分类里挑最接近的几条,字段以各自详情页为准
MCP 服务器工具数传输方式部署方式安装命令最后核验
Dbx MCP Server (dbx-mcp-server-9e27137c)当前 7 stdio · 本地进程 本地部署 有 2026-05-17
Git 838e2ec1 12 stdio · 本地进程 本地部署 有 2026-05-17
Markitdown MCP 2c653a5f 1 stdio · 本地进程 本地部署 有 2026-05-17
JavaGuide 未标注 stdio · 本地进程 本地部署 有 2026-04-11
awesome-claude-skills 未标注 stdio · 本地进程 本地部署 有 2026-04-26
Context7 831acf40 未标注 stdio · 本地进程 本地部署 有 2026-05-17

服务器档案

开发工具
字段来自项目仓库与公开索引,中文说明由本站标注
提供方ms-mfg-community上游标注的维护方 ms-mfg-community
分类开发工具本站按上游分类展示
连接方式stdio · 本地进程客户端在本机拉起子进程,数据不出本机
鉴权方式Token远程服务建议用最小权限凭据
维护方社区 / 第三方核验等级 仅资料核对
包名 / 运行时未标注 · Python本地安装时用到的包标识与运行时
部署类型本地部署上游给出的部署形态
最近核验2026-05-17上游最后核验时间

详细介绍

10.3 KB
来自项目仓库与上游目录的原文,未做删改

Databricks Error Logs MCP Server

An MCP (Model Context Protocol) server that enables GitHub Copilot to search and analyze application error logs stored in Databricks. Developers can get error context directly in their editor.

Architecture```mermaid

graph LR
subgraph "Option A: Azure Hosted"
IDE_A["🖥️ VS Code"] -- "HTTPS + Sub Key" --> APIM["🔐 APIM"]
APIM -- "HTTPS" --> CA["🐳 Container App"]
CA -- "SQL via PAT" --> DBX_A["🔶 Databricks"]
end

subgraph "Option B: Local HTTP"
    IDE_B["🖥️ VS Code"] -- "HTTP + Headers" --> LOCAL["🖥️ localhost:8000"]
    LOCAL -- "SQL via PAT" --> DBX_B["🔶 Databricks"]
end

subgraph "Option C: Local stdio"
    IDE_C["🖥️ VS Code"] -- "stdio" --> PROC["⚡ Local Process"]
    PROC -- "SQL via PAT" --> DBX_C["🔶 Databricks"]
end

> Full Azure architecture diagram and deployment details in AZURE_DEPLOYMENT.md

## Features

- **Search error logs** by error code, severity, file path, or message content
- **Get error frequency** — identify most common errors across your application
- **File-specific errors** — see all errors from a specific application file
- **Time-range search** — find errors within the last N hours
- **Severity summary** — quick overview of errors by severity level
- **Dynamic configuration** — specify Databricks workspace, warehouse, catalog at connection time
- **Three run modes** — stdio for local dev, local HTTP server for teams without Azure, streamable-http for Azure hosting

## Configuration

The server resolves Databricks connection details using this priority order:

1. **HTTP headers** (per-request, for remote/Azure hosting)
2. **Session config** (via the `configure_databricks` tool)
3. **Environment variables** (fallback, for local dev)

### HTTP Headers

When running in streamable-http mode (Azure), clients send config via headers:

| Header | Description | Example |
|--------|-------------|---------|
| `X-Databricks-Host` | Workspace URL | `https://adb-xxx.azuredatabricks.net` |
| `X-Databricks-Token` | Personal Access Token | `dapi...` |
| `X-Databricks-Warehouse-Id` | SQL Warehouse ID | `3b344f0c124e23fa` |
| `X-Databricks-Catalog` | Unity Catalog name | `dbx_1` |
| `X-Databricks-Schema` | Schema name | `default` |

### Environment Variables

For local development, set these in a `.env` file (copy from `.env.example`):```
DATABRICKS_HOST=https://adb-xxxxx.azuredatabricks.net
DATABRICKS_TOKEN=dapi...
DATABRICKS_WAREHOUSE_ID=xxxxx
DATABRICKS_CATALOG=dbx_1
DATABRICKS_SCHEMA=default
MCP_TRANSPORT=stdio
MCP_SERVER_PORT=8000

Session-Based Config (configure_databricks tool)

For clients that can't set custom HTTP headers, call the configure_databricks tool at the start of your session to set connection details.

Prerequisites

  1. Databricks Workspace with SQL warehouse configured
  2. Error logs parsed using the parse_error_logs.py notebook
  3. Databricks credentials: workspace URL, PAT, SQL Warehouse ID

Quick Start (Local)

stdio (simplest — VS Code manages the process)```bash

cd mcp_server
cp .env.example .env # Edit with your credentials
uv sync
uv run python -m databricks_error_logs_mcp


### Local HTTP Server (no Azure required)

For teams that **cannot deploy to Azure** — run the MCP server as a local HTTP service:```bash
cd mcp_server
cp .env.example .env   # Edit with your Databricks credentials

# Set transport to HTTP (PowerShell)
$env:MCP_TRANSPORT="streamable-http"

# Or on macOS/Linux:
# export MCP_TRANSPORT=streamable-http

uv sync
uv run python -m databricks_error_logs_mcp
# Server running at http://localhost:8000

> Why not Flask? MCP's streaming protocol requires ASGI. The server uses Starlette + uvicorn — Python's standard ASGI stack with an identical developer experience to Flask.

Test with the MCP Inspector:```bash
uv run mcp dev databricks_error_logs_mcp/server.py


## Azure Deployment

See AZURE_DEPLOYMENT.md for one-click `azd up` deployment behind Azure API Management.

## VS Code Integration

### Option A: Remote Server (Azure-hosted)

If the MCP server is deployed to Azure (see AZURE_DEPLOYMENT.md), configure VS Code to connect to the remote endpoint.

1. Open your workspace in VS Code
2. Create or edit `.vscode/mcp.json`
3. Add the following configuration:```json
{
  "servers": {
    "databricksErrorLogs": {
      "type": "http",
      "url": "https://<your-apim-name>.azure-api.net/mcp",
      "headers": {
        "Ocp-Apim-Subscription-Key": "${input:apim-key}",
        "X-Databricks-Host": "${input:databricks-host}",
        "X-Databricks-Token": "${input:databricks-token}",
        "X-Databricks-Warehouse-Id": "${input:databricks-warehouse-id}",
        "X-Databricks-Catalog": "${input:databricks-catalog}",
        "X-Databricks-Schema": "${input:databricks-schema}"
      }
    }
  },
  "inputs": [
    {
      "type": "promptString",
      "id": "apim-key",
      "description": "APIM Subscription Key",
      "password": true
    },
    {
      "type": "promptString",
      "id": "databricks-host",
      "description": "Databricks workspace URL (e.g. https://adb-xxx.azuredatabricks.net)"
    },
    {
      "type": "promptString",
      "id": "databricks-token",
      "description": "Databricks Personal Access Token",
      "password": true
    },
    {
      "type": "promptString",
      "id": "databricks-warehouse-id",
      "description": "Databricks SQL Warehouse ID"
    },
    {
      "type": "promptString",
      "id": "databricks-catalog",
      "description": "Unity Catalog name",
      "default": "dbx_1"
    },
    {
      "type": "promptString",
      "id": "databricks-schema",
      "description": "Schema name",
      "default": "default"
    }
  ]
}
  1. Replace ` with your APIM instance name (runazd env get-valuesto findAPIM_MCP_ENDPOINT`)
  2. Restart VS Code — Copilot will prompt you for each value on first use

> Note: "type": "http" tells VS Code to use the streamable-http MCP transport, which is the current standard for remote MCP servers.

Option B: Local HTTP Server (no Azure required)

Run the server locally as an HTTP service and connect VS Code over HTTP:

  1. Start the server locally (see Local HTTP Server Quick Start above)
  2. Create or edit .vscode/mcp.json:```json
    {
    "servers": {
    "databricksErrorLogs": {
    "type": "http",
    "url": "http://localhost:8000/mcp",
    "headers": {
    "X-Databricks-Host": "${input:databricks-host}",
    "X-Databricks-Token": "${input:databricks-token}",
    "X-Databricks-Warehouse-Id": "${input:databricks-warehouse-id}",
    "X-Databricks-Catalog": "${input:databricks-catalog}",
    "X-Databricks-Schema": "${input:databricks-schema}"
    }
    }
    },
    "inputs": [
    {
    "type": "promptString",
    "id": "databricks-host",
    "description": "Databricks workspace URL (e.g. https://adb-xxx.azuredatabricks.net)"
    },
    {
    "type": "promptString",
    "id": "databricks-token",
    "description": "Databricks Personal Access Token",
    "password": true
    },
    {
    "type": "promptString",
    "id": "databricks-warehouse-id",
    "description": "Databricks SQL Warehouse ID"
    },
    {
    "type": "promptString",
    "id": "databricks-catalog",
    "description": "Unity Catalog name",
    "default": "dbx_1"
    },
    {
    "type": "promptString",
    "id": "databricks-schema",
    "description": "Schema name",
    "default": "default"
    }
    ]
    }

3. Restart VS Code — Copilot will prompt for your Databricks credentials

> **Note:** No APIM subscription key needed — you're connecting directly to the local server. Replace `localhost` with a hostname/IP to share with teammates on your network.

### Option C: Local Server (stdio)

If running the MCP server locally on your machine:

1. Install uv
2. Navigate to the `mcp_server/` directory and run `uv sync`
3. Create or edit `.vscode/mcp.json`:```json
{
  "servers": {
    "databricksErrorLogs": {
      "type": "stdio",
      "command": "uv",
      "args": ["run", "python", "-m", "databricks_error_logs_mcp"],
      "cwd": "${workspaceFolder}/mcp_server",
      "env": {
        "DATABRICKS_HOST": "${input:databricks-host}",
        "DATABRICKS_TOKEN": "${input:databricks-token}",
        "DATABRICKS_WAREHOUSE_ID": "${input:databricks-warehouse-id}",
        "DATABRICKS_CATALOG": "${input:databricks-catalog}",
        "DATABRICKS_SCHEMA": "${input:databricks-schema}"
      }
    }
  },
  "inputs": [
    {
      "type": "promptString",
      "id": "databricks-host",
      "description": "Databricks workspace URL (e.g. https://adb-xxx.azuredatabricks.net)"
    },
    {
      "type": "promptString",
      "id": "databricks-token",
      "description": "Databricks Personal Access Token",
      "password": true
    },
    {
      "type": "promptString",
      "id": "databricks-warehouse-id",
      "description": "Databricks SQL Warehouse ID"
    },
    {
      "type": "promptString",
      "id": "databricks-catalog",
      "description": "Unity Catalog name",
      "default": "dbx_1"
    },
    {
      "type": "promptString",
      "id": "databricks-schema",
      "description": "Schema name",
      "default": "default"
    }
  ]
}
  1. Restart VS Code — Copilot will prompt for your Databricks credentials

Verify It Works

After configuring either option, open GitHub Copilot chat in VS Code and ask:

> "What are the top 10 most common errors?"

Copilot should invoke the get_error_frequency tool and return results from your Databricks workspace.

Available Tools

configure_databricks

Set Databricks connection details for the current session.

search_error_logs

Search for errors using multiple criteria (error code, severity, file path, message).

get_error_frequency

Get most frequently occurring errors with counts.

get_severity_summary

Overview of errors grouped by severity level.

get_file_errors

All errors from a specific application file.

search_by_message

Full-text search in error messages.

search_by_time_range

Find errors within the last N hours.

Development```bash

uv run pytest # Run tests
uv run ruff check . # Lint
uv run mcp dev databricks_error_logs_mcp/server.py # Inspector


## Docker```bash
docker build -t mcp-server .
docker run -p 8000:8000 mcp-server

The container defaults to streamable-http transport on port 8000.

常见问题

8 问
答案由本站依据条目字段整理,不引用上游原文
Dbx MCP Server (dbx-mcp-server-9e27137c) 怎么安装?

在客户端按下面的安装命令接入:uv run python -m databricks_error_logs_mcp。装完在客户端配置里引用同一个包或端点。

Dbx MCP Server (dbx-mcp-server-9e27137c) 怎么在 VS Code 里配置?

它的传输方式是stdio · 本地进程,在客户端的 MCP 配置里填好启动命令和参数,保存后完全退出客户端再打开,配置才会生效。

Dbx MCP Server (dbx-mcp-server-9e27137c) 收费吗?

本站不代收费用,也没有标注具体价格。是否收费取决于服务本身和背后的接口额度,正式使用前按官方说明确认。

Dbx MCP Server (dbx-mcp-server-9e27137c) 需要 API Key 吗?

资料标注需要Token认证,使用前准备好对应凭据,并按最小权限配置。

Dbx MCP Server (dbx-mcp-server-9e27137c) 支持哪些客户端?

资料标注支持 VS Code;其他客户端只要能配置 MCP Server,也能用同样的命令或地址接入。

Dbx MCP Server (dbx-mcp-server-9e27137c) 能提供哪些能力?

资料显示约 7 个工具;能力标签包括 开发工具、Python、VS Code、日志管理。具体工具清单以官方文档为准。

Dbx MCP Server (dbx-mcp-server-9e27137c) 需要什么运行环境?

运行时为 Python;属于本地进程,需要在本机装好运行环境。

Dbx MCP Server (dbx-mcp-server-9e27137c) 是本地还是远程?

它是本地进程(stdio),由客户端在本机拉起。传输方式为stdio · 本地进程。

热门专题

按标签挑服务器
全部专题 →

服务器信息以作者仓库与官方文档为准;分类、传输方式与鉴权方式的中文说明为中文整理,最近核验 2026-05-17。