Decide where the work belongs
Map loss consequence
For each case, record response timing, durable receipt, processing state, downstream effect, event age, receiver reach, and retry outcome. Separate a display update from a state transition that cannot be reconstructed safely after an interruption.
Compare response strategies
Assess recording before response, durable queue handoff, bounded synchronous work, explicit failure, or reconciliation of an uncertain event. Record capacity, idempotency, owner, and what a sender retry would do. Do not equate a fast response with a completed effect.
Set a queue order
Put confirmed missing receipts and shared early-response paths first, then isolated slow processing and reporting gaps. Add a fixture, review date, and trigger for new uncertain events. Keep ranking provisional when process lifetime or sender retries are unobserved.
What to carry forward
Return an acknowledgement queue with loss consequence, affected scope, safe containment, owner, and proof of durable handoff. Escalate events that may have been accepted without a recovery handle. Keep rank provisional when response or retry history is incomplete.
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