Keep the lesson for the next incident
Record response semantics
Document response conditions, durable receipt, queue handoff, interruption behavior, effect identity, duplicate handling, retry treatment, and owner. State which response statuses leave an event reviewable and which guarantee only receipt.
Keep crash fixtures
Retain crash-before-receipt, receipt-before-crash, handoff, slow, completed, duplicate, and retry cases with expected recovery and effect. Include the original ordering gap and remove payload values that do not affect durability. Assign an explicit recovery owner.
Review acknowledgement signals
Watch success responses without receipts, interrupted workers, uncertain effects, duplicate reviews, and retry mismatches by endpoint and receiver revision. Assign thresholds and an owner. Re-run the fixture set after response, queue, or storage changes. Review by receiver revision.
What to carry forward
Close learning with response semantics, interruption fixtures, recovery ownership, recurrence signals, and explicit sender limits. The useful outcome is an observable durable handoff, not a faster success response. Keep receipt, effect, and retry uncertainty attached to future reviews.
Technical background: Stripe documentation.
Keep the decision with the work.
Use a Work Item in Aglet to record the problem, the evidence you have, and the next decision. Add an owner and priority, then keep updates in the discussion so the next person can follow the reasoning.
Create an account See the product workflow