Check whether the outcome improved
Define acknowledgement outcomes
Write expected response, durable receipt, queue or worker handoff, process state, effect, and retry treatment for each interruption and success fixture. State which response is allowed when no recovery record exists.
Exercise crash boundaries
Interrupt processing before and after receipt, then inspect event identity and downstream state. Confirm a pre-receipt failure remains retryable or reviewable and a post-receipt retry cannot create a second effect.
Check slow and duplicate work
Run a slow but recoverable event and a duplicate after durable receipt. Verify response timing, receipt visibility, worker recovery, and effect count remain explainable. Record any sender behavior the fixture cannot prove. Compare receipt persistence after each restart.
What to carry forward
Accept when response implies a durable recovery handle, interruptions remain visible, and retries preserve identity and effect safety. Keep verification open for unobserved sender retry policy. Record interruption fixtures, receiver revision, response boundary, and remaining uncertainty. Record any sender retry limit.
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