cdp-mcp-server
MCP server for Cloudera Manager and CDP cluster administration, enabling monitoring, configuration, and troubleshooting of Hadoop ecosystem services through natural language.
README
cdp-mcp-server
MCP (Model Context Protocol) server for Cloudera Manager / CDP cluster administration and troubleshooting.
Fork of dvergari/cloudera-mcp-server — extended with pluggable registry backends (File, Env, Iceberg) and additional service clients (YARN, Spark History Server, HDFS NameNode, Oozie).
Quick Start (general purpose — no Iceberg required)
Option 1: FileRegistry (recommended for development)
# Install
git clone https://github.com/disoardi/cdp-mcp-server.git
cd cdp-mcp-server
python3.12 -m venv .venv
source .venv/bin/activate
pip install -e .
# Configure
cp cm_instances.yaml.example cm_instances.yaml
# Edit cm_instances.yaml with your CM credentials
# Run
REGISTRY_BACKEND=file cdp-mcp
Option 2: EnvRegistry (single CM, zero config)
REGISTRY_BACKEND=env \
CM_HOST=cm.example.com \
CM_USERNAME=admin \
CM_PASSWORD=changeme \
CM_USE_TLS=false \
cdp-mcp
Claude Desktop Configuration
FileRegistry
{
"mcpServers": {
"cdp": {
"command": "/path/to/cdp-mcp-server/.venv/bin/cdp-mcp",
"env": {
"REGISTRY_BACKEND": "file",
"REGISTRY_FILE_PATH": "/path/to/cm_instances.yaml"
}
}
}
}
EnvRegistry (single CM)
{
"mcpServers": {
"cdp": {
"command": "/path/to/cdp-mcp-server/.venv/bin/cdp-mcp",
"env": {
"REGISTRY_BACKEND": "env",
"CM_HOST": "cm.example.com",
"CM_USERNAME": "admin",
"CM_PASSWORD": "changeme",
"CM_USE_TLS": "false"
}
}
}
}
Available Tools
Cloudera Manager (original dvergari tools)
list_clusters— List all managed DataHub clusterslist_services— List services on a clusterget_service_logs— Extract service logs with time range filteringget_alerts— Get cluster alerts and eventsget_service_metrics— Time-series metrics via tsqueryget_config/update_config— Read and write service configurationrun_service_command/get_command_status— Execute async CM commandsget_host_status— Host health and role inventoryget_audit_events— CM audit loglist_datahubs— Enumerate DataHub clustersrefresh_cluster_map— Rebuild cluster→CM mapping
Registry Management
registry_list— List registered CM instancesregistry_stats— Registry statisticsregistry_add— Register a new CM instanceregistry_deactivate— Deactivate a CM instanceregistry_update_field— Update a registry fieldregistry_reload— Hot-reload registry
YARN ResourceManager
get_yarn_app— Get YARN application details and diagnosticslist_yarn_apps— List YARN applications with filtersget_yarn_queue— Get YARN scheduler queue capacity/usage
Spark History Server
get_spark_app— Get Spark application summarylist_spark_apps— List Spark applicationsget_spark_stages— Get stage details (useful for debugging slow jobs)
HDFS NameNode
get_namenode_status— NameNode health, capacity, corrupt/missing blocks
Oozie
get_oozie_job— Get workflow or coordinator job detailslist_oozie_jobs— List Oozie jobs with filters
Registry Backends
| Backend | Use case | Requires |
|---|---|---|
file |
Development, small teams | cm_instances.yaml |
env |
Single CM, quick tests | env vars CM_HOST, CM_USERNAME, CM_PASSWORD |
iceberg |
Production CDP environments | Impala/HiveServer2 + Iceberg table |
License
Apache License 2.0 — see LICENSE. Original work © Davide Vergari. Modifications © Davide Isoardi.
推荐服务器
Baidu Map
百度地图核心API现已全面兼容MCP协议,是国内首家兼容MCP协议的地图服务商。
Playwright MCP Server
一个模型上下文协议服务器,它使大型语言模型能够通过结构化的可访问性快照与网页进行交互,而无需视觉模型或屏幕截图。
Magic Component Platform (MCP)
一个由人工智能驱动的工具,可以从自然语言描述生成现代化的用户界面组件,并与流行的集成开发环境(IDE)集成,从而简化用户界面开发流程。
Audiense Insights MCP Server
通过模型上下文协议启用与 Audiense Insights 账户的交互,从而促进营销洞察和受众数据的提取和分析,包括人口统计信息、行为和影响者互动。
VeyraX
一个单一的 MCP 工具,连接你所有喜爱的工具:Gmail、日历以及其他 40 多个工具。
graphlit-mcp-server
模型上下文协议 (MCP) 服务器实现了 MCP 客户端与 Graphlit 服务之间的集成。 除了网络爬取之外,还可以将任何内容(从 Slack 到 Gmail 再到播客订阅源)导入到 Graphlit 项目中,然后从 MCP 客户端检索相关内容。
Kagi MCP Server
一个 MCP 服务器,集成了 Kagi 搜索功能和 Claude AI,使 Claude 能够在回答需要最新信息的问题时执行实时网络搜索。
e2b-mcp-server
使用 MCP 通过 e2b 运行代码。
Neon MCP Server
用于与 Neon 管理 API 和数据库交互的 MCP 服务器
Exa MCP Server
模型上下文协议(MCP)服务器允许像 Claude 这样的 AI 助手使用 Exa AI 搜索 API 进行网络搜索。这种设置允许 AI 模型以安全和受控的方式获取实时的网络信息。