Decide where the work belongs
Identify repeated or abandoned work
For controlled examples, compare accepted records before and after retry. Record duplicates, unintended updates, and failed rows that were never retried. Keep each outcome tied to source identity so a changing total count does not conceal one record repeated while another remains unresolved.
Scope the partial-failure population
Group imports by the reason for partial completion and the retry path users choose. Compare validation failures with interruptions that leave durable outcomes uncertain. Estimate reach from these paths rather than all imports, and retain uncertainty when summary events do not reveal which records execution actually accepted.
Assess the available reconciliation
Try the outcome report and corrected-file flow with a known fixture. Determine whether users can identify exactly which records need another attempt and whether the next scope is visible before confirmation. Include manual source comparison and cleanup effort when the flow requires users to infer accepted state themselves.
What to carry forward
Record priority with confirmed repeated work, unresolved rows, affected failure paths, and recovery effort. Assign an escalation trigger for duplicate acceptance or an unknown outcome presented as safe to retry. Keep the decision focused on exact retry scope rather than adding a generic retry button to every failed import.
Technical background: Google Cloud 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