Build a useful investigation brief
Align due time with execution
Map schedule calculation, timezone and calendar interpretation, due selection, enqueue, queue acceptance, claim, worker start, lock, dependency, commit, response or output, retry, and final visibility. Compare intended and actual times. Record clock or scheduler evidence that is unavailable instead of guessing.
Test failure and overlap conditions
Reproduce a dependency failure, delayed claim, missed due window, or overlapping run with the same schedule and input structure in a controlled environment. Compare run identifiers, locks, attempts, side effects, and final state. Keep separate schedule identities to check isolation.
Evaluate one safe recovery change
Try one correction such as explicit timezone handling, durable due-run creation, an overlap guard, retry classification, or a reviewable missed-run state. Replay normal, failed, late, and overlapping cases. Check idempotency, final records, schedule continuity, and visible status before accepting.
What to carry forward
Deliver an evidence-backed brief with due-to-outcome timeline, calendar context, first failure boundary, overlap and retry behavior, tested correction, and uncertainty. Investigation is complete when implementation can target one schedule boundary and verification can cover normal, missed, late, and overlapping runs.
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