Decide where the work belongs
Rank the repeated effect
Name the record, transition, message, or external action produced twice and whether repetition changes money, ownership, notifications, or audit history. Separate harmless duplicate reads from irreversible or confusing writes. Give higher priority to effects whose correct winner cannot be identified.
Group reach by trigger and workflow
Compare repeated submissions, request timeouts, job retries, event replays, concurrency, route, account class, and release window. Distinguish confirmed duplicates from near misses prevented by a guard. A rare severe effect may outrank a common harmless repeat.
Choose a bounded deduplication slice
Select immediate investigation, an idempotency or uniqueness improvement, or collection of one missing attempt-to-effect trace. State what changes priority, such as duplicate durable records or repeated external effects. Keep broad data repair separate from the first operation that can be made safe.
What to carry forward
Produce a queue decision tied to effect risk, confirmed reach, trigger, and repair uncertainty. Include whether duplicates are new or longstanding. The result is actionable when the next engineer has one logical operation to reproduce and understands why its effect risk outranks other reliability work.
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