Maintainer sponsor for a blog idea: metrics for AI agent tools

[tridev-tej]

Following step 0 in the blog guide, I'm looking for a maintainer to discuss and sponsor a post on bounded metrics for AI agent tools.

Planned synthetic example: compare a fixed set of tool/outcome label values with per-run-ID labels, then report series counts and the test's limits. AI-assisted idea and outline; the example and measurements are unbuilt. I work on AI agents and backend systems.

Would this fit, and who could sponsor the technical review?

Tejesh Vaish

[ArthurSens]

Hi Tejesh!

For a Prometheus blog, it would be nice if the content covered the usage of our MCP server (if that makes sense with what you have in mind), so cc @prometheus/prometheus-mcp-maintainers. Would that be something you'd be happy to collaborate on?

[tridev-tej]

Thanks @ArthurSens, yes, happy to collaborate. One angle could be using prometheus-mcp to investigate label growth in a small agent-tool workload. Start with run IDs as labels, inspect the resulting series through MCP, then compare with a bounded set of tool/outcome labels. The walkthrough would include exact PromQL and measured outputs so readers can check the findings. Is that useful to the MCP maintainers, or is there another workflow you'd prefer?