Adrian Cole

fyi I proposed a clarification on MCP spec about how to encode trace identifiers in MCP's JSON-RPC requests (which I think would end up implemented here decoupled from MCP) https://github.com/modelcontextprotocol/modelcontextprotocol/pull/414...

I also ran into this today. I want to make a demo that uses embeddings and chat at the same time capable of working with ramalama. I think a simple...

So I have another [conference soon in NZ](https://www.cloudnativesummit.co/), so wondering if we're close on this. The use case is `ramalama serve my-chat-model my-embeddings-model` or just `serve` and allow already pulled...

@carlocab fyi this is needed for envoy AI gateway and without which there are functionality problems on macos. can you take a look?

@SMillerDev if you are like me and hacking OSS on a weekend, mind having a look? we're pretty jammed up until this PR is in sadly

I was thinking to ignore them by default and have an arg to override (possibly zero out) the skip URL pattern

agree it is a choice between annoyance and config, vs a chance someone will care about not seeing /health and then be annoyed at lack of config ;)

This is also interestingly an issue for streamable MCP as well as the HTTP POST spans end up as new traces (at least in openai-agent) https://github.com/elastic/observability-examples/tree/main/inference-platforms/aigw#mcp-agent Here's one of the...

yeah I think we can have something like that, a comma-separated list of HTTP paths to ignore, since almost always it is http

glad to see most of your http status are 200 ;) seriously, pretty neat. Is this aggregation on the same data that matches a query for traces?