UCloud

You run an MCP server that exposes UCloud instance management through the MCP protocol, enabling you to query instance information, list all instances, check status, and monitor metrics via a standardized interface. This server handles MCP protocol I/O and can be configured with UCloud credentials to interact with your cloud account securely.

How to use

Start the MCP server and connect an MCP client to the standard input/output interface. You can provide a custom configuration file or rely on environment variables for UCloud credentials. Use the MCP client to request: detailed instance information, a full instance list, real-time status, and monitoring metrics such as CPU, disk, and network usage. The server exposes endpoints that return structured data for these actions, enabling you to build dashboards or automate checks.

Typical usage pattern includes starting the server with a configuration file and optional port, then issuing MCP requests from your client to perform the supported operations. Responses are designed to be consumed by automation tooling or dashboards.

How to install

Prerequisites you need before installation: a Go toolchain and a UCloud account with API credentials.

Step-by-step commands to install and run the MCP server locally:

# Prerequisites
# Ensure Go is installed (version 1.23 or higher)

# Clone the MCP server repository
git clone https://github.com/renzheng.wang/ucloud-mcp-server.git
cd ucloud-mcp-server

# Install dependencies
go mod download

# Build the MCP server binary
go build -o ucloud-mcp-server

# Run with default configuration (reads ./config.json if present)
./ucloud-mcp-server

# Run with a custom config and port
./ucloud-mcp-server --config /path/to/config.json --port 8080

Configuration

Configure credentials and region for UCloud access. You can use a configuration file or environment variables. The configuration file should include your region, project_id, public_key, and private_key. If you omit values in the file, the server will read them from environment variables when available.

{
  "region": "cn-bj2",
  "project_id": "your-project-id",
  "public_key": "your-public-key",
  "private_key": "your-private-key"
}

Environment variables

If you prefer environment variables, set these values in your environment before starting the server.

export UCLOUD_REGION="cn-bj2"        # UCloud region
export UCLOUD_PROJECT_ID="your-project-id"  # Project ID
export UCLOUD_PUBLIC_KEY="your-public-key"  # API public key
export UCLOUD_PRIVATE_KEY="your-private-key"  # API private key

Available startup options

You can start the server with a configuration file and an optional port. The default port is 8080 and the default config path is ./config.json.

# Use custom config file
./ucloud-mcp-server --config /etc/ucloud/config.json

# Use custom port
./ucloud-mcp-server --port 9000

# Use both custom config and port
./ucloud-mcp-server --config /etc/ucloud/config.json --port 9000

Notes

Keep your UCloud API credentials secure. Use configuration files or a secret management service for production deployments. All operations are performed through the MCP protocol via standard I/O. Monitoring data may have a few minutes of delay, so design your dashboards with that latency in mind.

Security and monitoring considerations

Limit access to the server and protect credentials. Regularly rotate API keys and monitor usage. Use secure channels for any remote management and review monitoring metrics to catch anomalies early.

Troubleshooting

If the server fails to start, verify that your configuration file path is correct, the port is not in use, and your UCloud credentials are valid. Check that environment variables are exported when using the environment-based configuration.

Available tools

GetInstanceInfo

Fetch detailed information about a specific MCP instance, including basic details, configuration, current status, and resources.

ListInstances

Retrieve a list of all available MCP-managed instances with basic information and current status.

GetInstanceStatus

Query real-time operational status for a selected instance.

GetMetrics

Access monitoring metrics for instances, including CPU, disk, and network statistics.