Build a useful investigation brief
Build lifecycle fixtures
Create cases for intentional pause, failure disable, repaired configuration, first post-recovery delivery, pending event, duplicate replay, and an event whose effect is already known. Record expected status and disposition before execution.
Follow one event through recovery
Capture endpoint status, disable reason, configuration revision, re-enable time, delivery identity, response, durable receipt, downstream effect, and replay decision. Compare a working recovery with a failing one at the first divergent boundary.
Vary recovery timing
Change one configuration value, endpoint state, receiver revision, pending age, capacity condition, or replay choice. Compare delivery and effect without sending an unbounded backlog. If sender retention is unknown, state that missing evidence directly.
What to carry forward
The brief is ready when lifecycle fixtures identify the first recovery divergence and show pending and completed event consequences. Recommend the smallest safe status or reconciliation step. Leave replay safety unresolved when event identity or prior effect cannot be proven.
Technical background: Zendesk developer 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