alibabacloud-ack

Alibaba Cloud's ack-mcp-server unifies container operations capabilities, enabling AI assistants and third-party AI agents to perform complex tasks via natural language through the MCP protocol, empowering container-native AIOps. DingTalk discussion group: 70080006301

Installation

Run this command in your terminal to add the MCP server to Claude Code.

Run in terminal:

View docs

Command

claude mcp add --transport stdio aliyun-alibabacloud-ack-mcp-server python -m main_server --transport sse --host 0.0.0.0 --port 8000 --allow-write \
  --env REGION_ID="your-region-id" \
  --env ACCESS_KEY_ID="your-access-key-id" \
  --env KUBECONFIG_MODE="ACK_PRIVATE" \
  --env ACCESS_KEY_SECRET="your-access-key-secret"

How to use

ack-mcp-server exposes a unified MCP interface for Alibaba Cloud ACK and Kubernetes operations, enabling natural language powered interactions with the underlying cluster and observability tooling. Core capabilities include native Kubernetes operations via ack_kubectl, resource lifecycle actions for ACK clusters (list, describe, create, delete, upgrade for clusters and components are labeled as Later), and integrated observability features such as Prometheus metric queries, PromQL guidance, control plane log queries (SLS/SQL and NL-to-SLS-SQL), and audit log queries. The server supports several deployment options and can be connected to multiple AI agents and CLIs (e.g., Qwen Code, Qoder CLI, Gemini CLI, Claude Code, Codex). To use, run the server and configure your MCP client to point at the endpoint; then register the server with the agent or CLI you are using. Typical workflows include querying cluster health, inspecting resources, fetching logs, and running diagnosing routines across ACK clusters through natural language prompts that are translated into MCP actions.

How to install

Prerequisites:

Python 3.12+ installed on your machine or your deployment environment
Access to an ACK cluster and valid API credentials (AccessKeyId/AccessKeySecret)
Network access to the cluster API (as configured by kubeconfig and KUBECONFIG_MODE)

Install steps:

Clone the repository git clone https://github.com/aliyun/alibabacloud-ack-mcp-server cd alibabacloud-ack-mcp-server
Set up a Python virtual environment (recommended) python -m venv .venv source .venv/bin/activate
Install dependencies uv sync # or pip install -r requirements.txt
Prepare environment variables (example) cp .env.example .env vim .env

Ensure ACCESS_KEY_ID, ACCESS_KEY_SECRET, and other config values are set
Run the server (example using Makefile target if available) make run
Alternative deployment options provided in the README:
- Docker: pull and run the ACK MCP server image and pass credentials as env vars
- Binary: build or download prebuilt binaries and execute the main server entrypoint
- Helm: deploy into a Kubernetes cluster with Helm using the provided charts
Verify the server is reachable and functional via the configured endpoint and transport (e.g., SSE) and register with your MCP client or AI agent

Additional notes

Notes and tips:

Ensure your AccessKeyId/AccessKeySecret have the minimum required permissions (cs read-only, log read-only, arms read for Prometheus, etc.) as described in the project documentation.
If you use ACK_PRIVATE networking, ensure kubeconfig access is available to the MCP server and set KUBECONFIG_MODE accordingly.
When running in Docker or binary modes, expose the port used by the server (default 8000) and configure environment variables for credentials.
The server supports multiple agent integrations; consult the README for specific command lines to register each agent (Qwen Code, Qoder, Gemini, Claude Code, Codex).
If you encounter connectivity or permission errors, verify network access to the ACK API server and ensure the correct region is configured.
For development, using uv or a local environment is recommended to test MCP interactions before deploying to production.