Cover security and operational gates

Separate runtime and operator roles for OpenTelemetry traces across retrieve → rerank → generate for support answers. Runtime may only perform the narrow actions that on-call engineer debugging a slow groundedness miss needs; operators get audited break-glass with TTL. Encode a negative probe that denies the privilege trick related to missing parent context → orphan spans, cannot join retrieve to answer.

Redact prompts/PII at collection. Secrets enter via a manager or workload identity — never source, fixtures, or exception strings. Incident TRACE-ORPHAN-441 should be impossible if these controls hold. Output allowlists and schema checks stay in force on error paths.

Rehearse the rollback or kill switch timed against a clock. Record actor, reason, prior revision/secret/flag, and verification query. Invariant reminder: every span carries trace_id + redacted attrs; prompts sampled ≤ 1%.

Abuse cases unique to this lab include the privilege path implied by missing parent context → orphan spans, cannot join retrieve to answer. Prove a read-only role cannot mutate. Break-glass tokens expire; leftover tokens fail the drill. Dependency pin/digest story matters when images or models move under you. Document how to rotate the credential that OpenTelemetry traces across retrieve → rerank → generate for support answers uses without a full outage window longer than your dual-run plan. Security evidence is part of ship, not an appendix nobody reads.

List network egress destinations and justify each. Ensure debug endpoints are off by default in the shipping config. Verify that error responses do not echo secrets or raw stack frames to clients. For multi-tenant paths, add a cross-tenant probe fixture. Time the rollback drill twice — once with the author, once with a peer. Store the drill transcript beside the threat notes for the incident id. In this chapter the product is OpenTelemetry traces across retrieve → rerank → generate for support answers, the human stakeholder is on-call engineer debugging a slow groundedness miss, and the incident id you design against is TRACE-ORPHAN-441. Re-state the oracle in your notes — fixture question produces 3 child spans; total latency attributes sum within 5% of root — and keep the invariant visible: every span carries trace_id + redacted attrs; prompts sampled ≤ 1%. Track trace_completeness ≥ 0.99 and p95_end_to_end_ms ≤ 2500 as the scoreboard. Surface under change control: POST /v1/support/answer. If you only have forty minutes, finish the fixture for missing parent context → orphan spans, cannot join retrieve to answer before polishing UI. Promotion language stays ternary: promote, hold, or roll back based on evidence, not hope.