Prometheus metrics

[jpinsonneau]

Having prometheus to monitor kubernetes / openshift resources is very common.

I wonder if that MCP server would be the right place implement tools to query prometheus metrics / alerts ?
That could be done in a very generic way.

FYI, Grafana is developing its own mcp server but for cloud only currently.

Implementing an optionnal tool in this repo seems to be the most independent way to me right now and would avoid having yet another MCP server running in parallel.

Any thoughts ?

[ardaguclu]

In my opinion, Prometheus can be an another MCP server focusing solely on this purpose. Otherwise, it would be hard to maintain single repository that may require knowledge of various, different aspects.