Post

Starting to wonder if the next gen of o11y products isn’t just: ingests OTLP into ClickHouse and/or DataLake for you and allows you to set alerts (for agents, not people)

Kinda? I agree the UI part isn't the "most" useful part. Where the UI becomes useful is helping a human understand the data, and claude artifacts are "fine", but still not great. From a backend perspective, its more than just querying data with SQL/PromQL. The backend can do more.

We've been seeing this with people resorting to "codemode" type mechanisms. We've taken a different stance and our MCP does more than just an API reference, it does so much more than that, and does it server-side. Providing additional context, and rendering is something the backend can do.

I don't think MCP should be just an API reference so I agree there, but also I wouldn't call codemode something to "resort" to. It's a really important primitive. I'm also trying to think a year+ out. Claude artifacts are not great... yet.

I've spent the last few months building something (hopefully released in a few weeks) using fully agentic development. Obviously it has top notch observability What I've noticed is that codemode is a kludge used when the MCP can't do it. String together APIs to perform a higher order outcome

The "ideal" scenario is that the backend already provides the analysis you want, packages that up and send that to the agent. If that's fast enough, you get way better workflows with an agent. Time is important for synchronous coding with an agent, even in async to an extent.

So to me, codemode is covering a gap in the MCP (or API for that matter). Can the MCP bring back the most used HTTP routes on the pods with the most CPU? What about a breakdown of all the attributes unique to a subset of logs in a wider dataset?