Triage and recover incident response

Map failure classes for POST /v1/rag/answer: retryable vs fatal vs needs-human. Retries need budgets, jitter, and idempotency rules aligned to declare early; preserve evidence; mitigation scoped; no blame in channel. The chapter’s signature failure — team debugs for an hour without declaring — customers keep seeing wrong policy cites — must take a deliberate branch, not a generic catch-all.

Implement the degrade/rollback/refuse behavior incident commander coordinating mitigate → communicate → verify needs when INC-RAG-CIT-2201 repeats. Prefer scoped controls (one flag, one weight, one tenant, one secret version) over fleet-wide restarts. Preserve evidence; do not delete logs to “clean the demo.”

After containment, check time_to_declare_m, time_to_mitigate_m, customer_update_sent moves in the safe direction and watch for retry amplification. Write the stop condition that ends the incident response for this lab.

Chaos note: inject only one fault class at a time and restore fixtures after. Watch for dual failures — dependency down and retry amplifier — which is how team debugs for an hour without declaring — customers keep seeing wrong policy cites becomes an outage. Customer communication templates (even if only for the drill) beat silence. If you queue deferred work, define poison-message handling. Budget documents should state the maximum extra spend allowed during retries. Close the loop by linking the containment action to a dashboard panel for time_to_declare_m, time_to_mitigate_m, customer_update_sent.

Draw a state diagram for degrade modes and put it in the repo as ASCII if needed. Cap concurrent retries across the process, not only per request. Ensure cancellation propagates to downstream HTTP clients. When failing closed, choose a user-visible message that does not leak internals. Practice the single command that flips the kill switch or weight to zero. After recovery, drain or inspect deferred work before declaring green. In this chapter the product is correctness incident for invalid RAG citations (SEV2), the human stakeholder is incident commander coordinating mitigate → communicate → verify, and the incident id you design against is INC-RAG-CIT-2201. Re-state the oracle in your notes — INC-RAG-CIT-2201: feature flag off in 8m; customer update sent; groundedness restored — and keep the invariant visible: declare early; preserve evidence; mitigation scoped; no blame in channel. Track time_to_declare_m, time_to_mitigate_m, customer_update_sent as the scoreboard. Surface under change control: POST /v1/rag/answer. If you only have forty minutes, finish the fixture for team debugs for an hour without declaring — customers keep seeing wrong policy cites before polishing UI. Promotion language stays ternary: promote, hold, or roll back based on evidence, not hope.