Content plan

Outcome

Learners can instrument a system so production tells them the truth, and close the loop.

Planned coverage

• Production as the outermost rung of Module 3’s ladder — the loop that catches what no test imagined.
• Observability setup: logs, metrics, and traces.
• What to instrument: tie observability to the failure model, so “what must never happen” is watched in production too, not just in tests.
• The loop in practice: telemetry → detect → diagnose → fix → ship → observe again.
• Closing discovery ↔ delivery: telemetry feeds the next round of discovery.

Agent payoff

Rich observability is the substrate that makes AI-assisted debugging possible in Part 4. An agent can only investigate what you instrumented.

Seeds

The telemetry set up here is exactly what Part 4’s AI debugging will query, for example through MCP.