Server Status Monitoring Tool

This is a server status monitoring tool based on FastMCP, capable of fetching CPU, memory, and uptime information from local or remote servers.

Features

• Supports fetching the status of local servers
• Supports fetching the status of remote servers via SSH
• Automatically reads SSH configuration files (~/.ssh/config)
• Supports custom SSH connection parameters
• Can be run as an independent server
• Can be integrated into other applications as an npm package

Installation

Install via npm

bash

Global installation

npm install -g server-status-mcp-server

Or install as a project dependency

npm install server-status-mcp-server

Install from Source Code

bash

Clone the repository

git clone https://github.com/nnnnzs/server-status-mcp-server.git
cd server-status-mcp-server

Install dependencies

npm install

Build the project

npm run build

Usage

Running as a Standalone Service

1. Start the Server

bash

If the package is installed globally

server-status-mcp-server

Or start from source code

npm start

2. Run the Test Client

bash
node client.js

3. Direct Command Line Invocation

Communicate with the server using stdio:

bash
echo {"jsonrpc":"2.0","method":"get_server_status","params":{},"id":1} | node dist/index.js

Fetch the status of a remote server:

bash
echo {"jsonrpc":"2.0","method":"getRemoteServerStatus","params":{"host":"172.18.1.103"},"id":1} | node dist/index.js

Integrating as an npm Package

1. Create and Configure MCP Service

javascript
import { createServerStatusMCP } from 'server-status-mcp-server';

// Create an instance of the MCP service
const server = createServerStatusMCP({
// Optional: Custom SSH configuration file path
sshConfigPath: '/path/to/ssh/config'
});

// Start the service
server.start({
transportType: "stdio" // or other transport types
});

2. Invoke Using MCP Client

javascript
import { MCPClient } from 'fastmcp';
import { createServerStatusMCP } from 'server-status-mcp-server';
import { spawn } from 'child_process';

async function main() {
// Method 1: Create a child process to run the service
const serverProcess = spawn('node', ['path/to/your/server.js']);

// Create an MCP client
const client = new MCPClient();

// Connect to the service process
await client.connect({
process: serverProcess,
transportType: "stdio"
});

// Method 2: Use in the same process (not recommended for production)
const server = createServerStatusMCP();
server.start({ transportType: "memory" });

const memoryClient = new MCPClient();
await memoryClient.connect({ transportType: "memory", server });

// Call the tool
const localStatus = await client.invoke('getLocalServerStatus', {});
console.log('本地服务器状态: ', JSON.stringify(localStatus, null, 2));

const remoteStatus = await client.invoke('getRemoteServerStatus', {
host: 'your-ssh-host'
});
console.log('远程服务器状态: ', JSON.stringify(remoteStatus, null, 2));

// Disconnect
await client.disconnect();
serverProcess.kill();
}

main().catch(console.error);

SSH Configuration Example

Add the following configuration to your ~/.ssh/config file:

Host my-server
HostName 192.168.1.100
User username
Port 22
IdentityFile ~/.ssh/id_rsa

Then you can use the configured hostname to fetch the status:

bash
echo {"jsonrpc":"2.0","method":"getRemoteServerStatus","params":{"host":"my-server"},"id":1} | node dist/index.js

Data Format Returned

Local Server Status

json
{
"cpuCount": 8,
"cpuRel": 0.12,
"memory": 8589934592,
"memoryUsage": "45.32",
"uptime": 123456,
"type": "local"
}

Remote Server Status

json
{
"cpu": "Processor information...",
"memory": "Memory usage...",
"uptime": "System uptime...",
"type": "remote"
}

API Documentation

Main Exports

typescript
// Create an MCP service instance
createServerStatusMCP(config?: { sshConfigPath?: string }): FastMCP

// Get local CPU usage
getLocalCpuUsage(): number

// Parse SSH configuration file
parseSSHConfig(configContent: string, targetHost: string): Promise<Record<string, string> | null>

// Server status interface
interface RemoteServerStatus {
cpuRel: string;
memoryRel: string;
alarmInfo: string;
itemStatus: string;
cpu: string;
memory: string;
memoryUsage: string;
uptime: string;
type: 'remote' | 'local';
}## Error Handling

If the connection fails or there is an error executing a command, the following will be returned:

json
{
"error": "Error message"
}

Publishing to NPM

If you wish to publish this package to NPM yourself, follow these steps:

1. Update the version number in package.json
2. Run npm run build to ensure the build is successful
3. Run npm login to log into your NPM account
4. Run npm publish to publish the package
5. Optional: Use npm publish --access public to publish as a public package

License

ISC