Model Context Protocol
Replies: 1 comment
-
|
I think your question is best suited for the azure team on the architecture of the server they're building rather than here and how it would scale for a complex enterprise use case. Likely the answer is they haven't built for that yet and are laying down the groundwork to get there. Or the eks aws built mcp, similar thing. Most of these mcp servers being built by the major players are in their infancy for maturity for enterprise use. The increase in number of tools requires more and more context / data for the ai to work through every time to determine if it should use a tool available. Hence the recommendation. |
Beta Was this translation helpful? Give feedback.
0 replies
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Uh oh!
There was an error while loading. Please reload this page.
-
Pre-submission Checklist
Question Category
Your Question
I'm using https://github.com/Azure/mcp-kubernetes to connect to my kubernetes cluster.
The limitation is it accepts only one kubeconfig and the tool calls will be limited to one cluster.
We have 200+ clusters in production.
So, i searched for wrappers which could help in deploying the same server with different configs: https://github.com/lamemind/mcp-server-multiverse
Works fine for few clusters.
But i see a limitation on the general tool count/performance where a best practice is only 40 tools for an MCP server.
But my design will cause ( number of clusters ) x ( the number of tools in one mcp server) enabled in a single MCP server, and hence this is not a good practice.
How to handle this scenario?
Beta Was this translation helpful? Give feedback.
All reactions