[OSPP-2025] Integrate karmada-mcp-server into dashboard

[warjiang]

Karmada (Kubernetes Armada) is a Kubernetes management system that enables you to run cloud-native applications across multiple Kubernetes clusters and cloud platforms without modifying the applications. By leveraging Kubernetes-native APIs and advanced scheduling capabilities, Karmada delivers truly open, multi-cloud Kubernetes.

Since OpenAI introduced LLM(large language models), various domains have been exploring their practical applications. The MCP (Model Context Protocol), developed by Anthropic, is a standardized protocol designed to rapidly and uniformly integrate vertical domain capabilities into existing workflows. The Karmada community has also been experimenting with large model applications, such as developing Karmada-MCP-Server based on the MCP protocol, which allows MCP-compatible clients to manage multi-cluster environments through natural language.

However, the current implementation requires extensive user configuration and remains relatively complex. Given that MCP is a standard protocol, we propose integrating Karmada-MCP-Server into the Karmada Dashboard. This integration would provide users with out-of-the-box large model capabilities through a ChatUI interface, significantly enhancing cluster management efficiency.

[warjiang]

with @kasanatte, we are working on integrate karmada-mcp-server into standalone cli, which will be resued in karmada dashboard

• cli screenshot:
  
• chatbot example page:
  

[warjiang]

things which will be further discussed:

• which transport is better, sse or stdio
• config for LLM feature
• multi provider support(including deepseek、qwen、gemini, currently we use openai-4o-mini)
• [optional] embedding the whole karmada website for further RAG, which will be used to generate karmada related yaml resources etc.
• [optional] multi cluster diagnosis which we still need more discussion. some related issue: Ability to view ResourceBindings on Karmada Dashboard #227