Check whether the outcome improved
Define lifecycle outcomes
Write expected endpoint status, configuration revision, delivery, response, pending disposition, receipt, effect, and replay decision for each recovery fixture. State which old events remain unknown and therefore cannot be replayed automatically.
Exercise bounded recovery
Repair one endpoint, re-enable it, and process a small known event set. Compare new, pending, duplicate, and completed identities. Confirm a repaired receiver does not receive an uncontrolled backlog and an existing effect remains protected.
Check pause and failure paths
Repeat an intentional pause and a failed re-enable, then inspect status and pending visibility. Verify the UI or record distinguishes paused, disabled, delivered, pending, replayed, and completed states for the reviewer. Compare by endpoint and environment.
What to carry forward
Accept when repaired endpoints deliver new work, pending state is visible, and bounded replay or closure preserves event identity and effect safety. Keep verification open for unobserved sender backlog rules. Record endpoint, fixture set, capacity boundary, and residual uncertainty.
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