Aglet

Prioritize webhook dead-letter recovery

Dead-letter work should prioritize records that can safely restore essential work before records that only need better labels. Unknown effects and expiring evidence require early attention. Weigh identity, failure history, retention, replay authority, and downstream capacity before choosing replay. Keep replay authority explicit.

Decide where the work belongs

  1. Map recovery value

    For each record, list event identity, resource, failure, attempt count, age, receipt, effect state, and next action. Separate a lost receipt from a failed side effect. Keep the payload protected while preserving safe correlation.

  2. Compare review paths

    Assess enrich-only review, current-state readback, bounded replay, closure, or owner escalation. Record authority, deduplication, and capacity requirements. Do not replay an item whose prior effect or identity cannot be established.

  3. Set the dead-letter order

    Queue unknown effects, expiring records, and essential work first, then visible but low-consequence failures. Set owner, review window, replay limit, and evidence threshold. Keep a large backlog below a smaller item that can cause a duplicate mutation.

What to carry forward

Return a dead-letter queue with identity, failure, age, effect state, recovery path, owner, and proof needed to reprioritize. Escalate unknown or expiring items. Keep ranking provisional when sender history or local effects are missing. Keep the review owner attached to each item.

Technical background: Twilio 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