Data Platform MCP Server
Provides read-only ETL log search using Grafana Loki query_range.
Provides read-only ETL log search across OpenSearch indexes.
Provides read-only catalog and SQL access to PostgreSQL, including listing schemas, tables/views, column metadata, table statistics, and read-only query explain plans.
Click on "Install Server".
Wait a few minutes for the server to deploy. Once ready, it will show a "Started" state.
In the chat, type
@followed by the MCP server name and your instructions, e.g., "@Data Platform MCP Serverwhat tables are in the analytics schema?"
That's it! The server will respond to your query, and you can continue using it as needed.
Here is a step-by-step guide with screenshots.
Data Platform MCP Server
目前版本 / Current release: v0.2.0
繁體中文:這是一個面向 Data Engineering / DataOps 的 Model Context Protocol (MCP) Server 參考專案,讓 ChatGPT、Claude、Codex 或其他 MCP Host 能以標準 MCP Tool、Resource、Prompt 介面安全地查詢資料平台。
English: A production-oriented Model Context Protocol (MCP) server reference for Data Engineering and DataOps. It exposes data-platform capabilities through standardized MCP tools, resources, and prompts.
繁體中文:Repository 預設只使用 synthetic demo data,不包含任何公司、客戶、真實資料庫帳密、內部 URL 或正式環境資訊。
English: The repository defaults to synthetic demo data and contains no company/customer data, real credentials, internal URLs, or production environment information.
v0.2 重點 / v0.2 Highlights
Official MCP Python SDK v2 (
MCPServer)stdio and Streamable HTTP transports
11 MCP tools for catalog, SQL, DataOps, logs, and runbooks
MCP resources + resource templates
MCP prompts for incident triage and data discovery
Synthetic zero-dependency demo mode
Optional read-only PostgreSQL catalog adapter
Airflow 3 stable public REST API (
/api/v2) adapterOpenSearch ETL log search
Grafana Loki ETL log search
SQL read-only guardrails
Docker / Docker Compose
Protocol-level MCP tests + pytest + Ruff CI
pip-audit + Trivy security workflow
Traditional Chinese / English documentation
Related MCP server: DataLakeHouseMCP
架構 / Architecture
ChatGPT / Claude / Codex / MCP Host
|
MCP (stdio / HTTP)
|
v
+---------------------------+
| Data Platform MCP Server |
| MCPServer v2 |
+-------------+-------------+
|
+-----------+-------------+----------------+
| | |
v v v
Catalog Adapter Operations Adapter Log Adapter
Demo / PostgreSQL Demo / Airflow 3 Demo/OpenSearch/Loki
| | |
+-------------------------+----------------+
|
v
Runbook KnowledgeMCP Primitives
Tools
Tool | Purpose / 用途 |
| Server health and version / 服務健康狀態 |
| List configured sources / 列出資料來源 |
| List schemas / 列出 Schema |
| List tables/views / 列出 Table/View |
| Column metadata / 欄位資訊 |
| Lightweight statistics / 輕量統計 |
| Read-only SQL explain / 唯讀 SQL 執行計畫 |
| List Airflow/demo DAGs / DAG 清單 |
| Latest DAG state / DAG 狀態 |
| Demo/OpenSearch/Loki log search / ETL Log 搜尋 |
| Operations knowledge / Runbook 搜尋 |
Resources
platform://capabilitiescatalog://{source}/{schema}/{table}
Prompts
incident_triage(dag_id, symptom)data_discovery(source, schema, question)
快速開始 / Quick Start
git clone https://github.com/kewinall/data-platform-mcp-server.git
cd data-platform-mcp-server
python -m venv .venv
source .venv/bin/activate
pip install -e '.[dev]'
cp .env.example .envstdio:
DPMCP_TRANSPORT=stdio data-platform-mcpStreamable HTTP:
DPMCP_TRANSPORT=streamable-http data-platform-mcpDefault MCP endpoint:
http://localhost:8000/mcpDocker:
cp .env.example .env
docker compose up --buildProduction Adapter Example
export DPMCP_MODE=postgres
export DPMCP_POSTGRES_DSN='postgresql://readonly_user:change-me@db:5432/analytics'
export DPMCP_OPERATIONS_MODE=airflow
export DPMCP_AIRFLOW_BASE_URL='https://airflow.example.internal'
export DPMCP_AIRFLOW_TOKEN='replace-me'
export DPMCP_LOGS_MODE=loki
export DPMCP_LOKI_URL='https://loki.example.internal'
export DPMCP_LOKI_TOKEN='replace-me'
data-platform-mcpSee docs/integrations.md.
Security Principles
All v0.2 platform integrations are read-only.
Use backend identities with minimum read permissions only.
PostgreSQL sessions enable
default_transaction_read_only=on.SQL write/DDL and multiple statements are rejected.
Airflow adapter uses GET endpoints only; it never triggers or clears DAGs.
OpenSearch adapter only searches indexes.
Loki adapter only uses
query_range.Secrets are supplied through environment variables and are not committed.
Application guardrails are defense-in-depth and do not replace backend-side least privilege.
Development
make install
make lint
make testCI validates Python 3.11/3.12, Ruff, pytest, compile, and Docker build. Security workflow runs pip-audit and Trivy.
Roadmap
v0.1 ✅ MCP foundation + PostgreSQL read-only adapter
v0.2 ✅ Airflow 3 + OpenSearch/Loki + MCP Resources/Prompts
v0.3 Vertica + metadata/lineage + parser-based SQL policy + RBAC/API token + audit
v0.4 Kubernetes/Helm + OIDC/multi-tenancy + OpenTelemetry + offline deployment
Documentation
docs/architecture.md— 架構 / Architecturedocs/tool-catalog.md— MCP Tool Catalogdocs/installation.md— 安裝 / Installationdocs/integrations.md— Airflow / OpenSearch / Lokidocs/security.md— Security Designdocs/v0.2.md— v0.2 Release Guidedocs/roadmap.md— RoadmapCHANGELOG.md— Release notes
License
MIT License. See LICENSE.
Tool Schema Changelog
Recent tool additions, removals, and schema changes observed during successful MCP inspections. Dates show when Glama detected each change.
No tool schema history has been recorded yet.
This server cannot be installed
Maintenance
Resources
Unclaimed servers have limited discoverability.
Looking for Admin?
If you are the server author, to access and configure the admin panel.
Related MCP Connectors
Let AI agents query data and act across all your business apps via MCP.
Query your warehouse or a CSV with Claude/ChatGPT over MCP, governed by table-level ACL + audit.
Governed data discovery, exact queries, decisions, simulations, and runtime utilities over MCP.
Query your org's data in natural language — read-only MCP access to SQL, NoSQL, files & warehouses.
Related MCP Servers
- AlicenseNot gradedqualityCmaintenanceEnables AI agents to query live schema, lineage, and query-context across data warehouses, dbt projects, orchestration systems, and BI tools via MCP tools.Apache 2.0
- FlicenseNot gradedqualityDmaintenanceEnables AI-powered MCP clients to interact with data lakehouse components including Kafka, Flink, and Trino/Iceberg for managing topics, jobs, catalogs, and executing queries.2-
- AlicenseNot gradedqualityCmaintenanceMCP server providing read-only Snowflake metadata tools (schemas, tables, queries, lineage) for agentic data pipeline generation, enabling natural-language-to-pipeline workflows with dbt, Airflow, and Great Expectations.1MIT
- AlicenseNot gradedqualityBmaintenanceEnables enterprise AI agents to query governed data lineage, PII-aware schema documentation, and semantic metadata from SQL logs via MCP, with role-based access and vector search.MIT
Latest Blog Posts
- Who's Calling? MCP Hosts Are an Identity Blind Spot (And the Spec Knows It)By Om-Shree-0709 on .mcpAgent IdentityOAuth 2.1
- Your AI Chatbot Just Exposed Your CEO's Salary to an InternBy Om-Shree-0709 on .Agent IdentityMCP SecurityOAuth Delegation
- Why MCP Servers Need Execution Sandboxing (And Why Your Current Stack Isn't Enough)By Om-Shree-0709 on .Agentic AiPrompt InjectionWebAssembly
MCP directory API
We provide all the information about MCP servers via our MCP API.
curl -X GET 'https://glama.ai/api/mcp/v1/servers/kewinall/data-platform-mcp-server'
If you have feedback or need assistance with the MCP directory API, please join our Discord server