Agent Observability: Hooks, Alloy, and Grafana _
We wired Claude Code and Codex into one Grafana stack with OpenTelemetry and Alloy, then used traces and logs to find and fix agent behavior issues at the source.
Free tier available. See plans →
> ls /var/log/decisions/*.md
Architecture decisions, MCP implementation notes, and the occasional rant about things that should not be this hard.
Latest dispatch
Start here if you want the freshest technical update before you drop into the archive.
We wired Claude Code and Codex into one Grafana stack with OpenTelemetry and Alloy, then used traces and logs to find and fix agent behavior issues at the source.
Archive
Architecture, observability, MCP surface design, and the patterns that survived contact with production.
Git commits become structured changelog entries and architectural decision records, then feed back into AI agents as queryable institutional memory.
We screened 3,500 skill candidates and adopted 466. A systematic mining and ingestion loop for building a coherent AI agent skill library at scale.
Agent systems are opaque by default. Progressive disclosure gives operators layered CLI views from quick status checks to full agent internals and decision traces.
AI agents hallucinate confidently. Ground truths are versioned, scoped facts that anchor agent behavior to reality. Here is how we built and enforce them.
One orchestrator routes work to specialist AI agents, each with distinct skills and memory. How Orkestra coordinates 46 agents and 466 skills in production.