SysOM MCP: Open-Source Intelligent O&M Assistant for AI-Powered System Diagnostics

This article introduces SysOM MCP, an open-source O&M assistant that enables AI Agents to perform automated system diagnostics via natural language using MCP.

By Ruiping Wan

AIOps New Paradigm: Perform O&M by Simply Speaking

Currently, operating system O&M faces challenges such as complex architectures, messy dependencies, difficulty in fault positioning, reliance on manual experience, fragmented tools, insufficient monitoring, and lack of automation. To address these issues, Alibaba Cloud has combined large language models (LLMs), Agents, and the Model Context Protocol (MCP) to achieve natural language-driven intelligent O&M: The LLM understands instructions, the Agent independently executes tasks, and MCP connects underlying diagnostic tools. The synergy among these three enables the AI assistant to automatically diagnose system problems, generating reports and repair suggestions, significantly improving efficiency and driving the evolution of O&M toward proactive intelligence.

The Alibaba Cloud Operating System Console (OS Console) is a one-stop operating system O&M management platform that provides powerful system diagnostic capabilities for memory, I/O, network, kernel crashes, etc. SysOM is the O&M component of the OS Console. However, these functions usually require users to log in to the console and possess certain O&M experience to be used effectively.

With the popularization of AI assistants (such as Qwen Code) users prefer to solve problems with a single sentence of natural language, such as "Why is my CPU usage so high?" To this end, SysOM has standardized and encapsulated its original diagnostic capabilities through MCP(Model Context Protocol) and launched the open-source project SysOM MCP. SysOM MCP evolved from the OS Console, transforming complex O&M operations into standard tools that AI can autonomously call, allowing the AI Agent to diagnose system problems "hands-on" like an engineer. Users don't need to understand commands—they simply ask questions in natural language to obtain precise system-level analysis. Today, SysOM MCP is driving natural language to become the new entry point for operating system diagnosis, making intelligent O&M truly inclusive and efficient.

SysOM MCP project open source address: https://github.com/alibaba/sysom_mcp

SysOM MCP: Driving System Diagnosis with Natural Language

Traditional O&M relies on command lines and expert experience, while general AI can "talk" but cannot "do." The emergence of SysOM MCP fills this gap—through the MCP protocol, AI can not only understand problems but also automatically execute real diagnostics, achieving a closed loop from "Q&A" to "action."

The SysOM MCP project includes over 20 production-grade diagnostic tools, all exposed through standard JSON-RPC over stdio/SSE, including:

• Memory diagnosis: Memory panoramic diagnosis, Java memory diagnosis, OOM memory diagnosis

• IO diagnosis: One-click IO diagnosis, IO traffic analysis diagnosis

• Network troubleshooting: Network packet loss diagnosis, network jitter diagnosis

• Scheduling diagnosis: System load diagnosis, scheduling jitter diagnosis

• Disk diagnosis: Disk analysis diagnosis

• System crash diagnosis: Crash diagnosis (dmesg analysis), crash diagnosis (deep vmcore analysis)

The project supports two modes: --stdio (local embedding) and --sse (HTTP service), easily integrating various AI clients.

To use SysOM MCP on an AI Agent platform that supports the MCP protocol (such as Qwen Code), first clone the project code locally:

git clone https://github.com/alibaba/sysom_mcp.git
cd sysom_mcp

Next, add the following configuration to the configuration file, allowing the AI assistant to drive operating system and O&M operations using natural language.

{
  "mcpServers": {
    "sysom_mcp": {
      "command": "uv",
      "args": ["run", "python", "sysom_main_mcp.py", "--stdio"],
      "env": {
        "ACCESS_KEY_ID": "your_access_key_id",
        "ACCESS_KEY_SECRET": "your_access_key_secret",
        "DASHSCOPE_API_KEY": "your_dashscope_api_key"
      },
      "cwd": "<sysom mcp project directory >",
      "timeout": 30000,
      "trust": false
    }
  }
}

Best Practice: Revealing Hidden Memory Leaks During Conversation

OS Copilot is an operating system intelligent assistant built by Alibaba Cloud based on large models. It supports natural language Q&A, auxiliary command execution, system O&M optimization and other functions, helping you use the Linux system more efficiently and effectively. Currently, OS Copilot on the OS Console has been connected to SysOM MCP. Users simply speak with OS Copilot in natural language to automatically trigger root cause troubleshooting of operating system problems. The diagnostic process requires no manual intervention, and results are clearly presented in a structured form, significantly lowering the O&M threshold and allowing complex problems to be "solved with a single question."

This article uses an inconspicuous memory leak as an example to demonstrate the diagnostic function of SysOM MCP.

The dialog in the figure above shows that OS Copilot gives possible causes for the leak. You can also click on the diagnostic report at the bottom of the figure for more detailed diagnostic results in the OS Console.

Download and Try it Today

SysOM MCP originated from the Alibaba Cloud OS Console, and the diagnostic tools have been verified in large-scale production environments.

Repository

The project uses a modern Python toolchain (uv + Python 3.11+), and installation is simple:

git clone https://github.com/alibaba/sysom_mcp.git
cd sysom_mcp && uv sync

Quick Start:

uv run python sysom_main_mcp.py --stdio # For local calls
uv run python sysom_main_mcp.py --sse --port 7140 # Start HTTP service

Use Cases

SysOM MCP can be connected to various AI agents, helping you build an intelligent assistant with system diagnostic capabilities.

Open Source Collaboration

🌟 GitHub address (Copy the link and open it in a browser): https://github.com/alibaba/sysom_mcp

Star, Fork, and contribute issues to help us build a new AI-native O&M ecosystem

Community

SysOM MCP: Open-Source Intelligent O&M Assistant for AI-Powered System Diagnostics

AIOps New Paradigm: Perform O&M by Simply Speaking

SysOM MCP: Driving System Diagnosis with Natural Language

Best Practice: Revealing Hidden Memory Leaks During Conversation

Download and Try it Today

Repository

Quick Start:

Use Cases

Open Source Collaboration

Read previous post:

OpenAnolis

You may also like

Comments

OpenAnolis

Related Products

Bastionhost

Managed Service for Grafana

AI Acceleration Solution

Offline Visual Intelligence Software Packages