# Make replay a controlled operation

Support needs to know what an event already did before trying it again. Hand over receipt history, effect state and the provider's redelivery behaviour together.

By Cobnex editorial. Published 2026-09-10. Updated 2026-09-11.

## Explain the processing states

Document accepted, pending, completed, failed and uncertain outcomes as used by the implementation. Show where event identity maps to the business operation.

Keep provider account scope visible in the authorised support tool. An event identifier without its source context can be ambiguous across integrations.

Provide a controlled way to inspect relevant payload details without exposing them in broad logs.

## Distinguish retry from replay

A worker retry continues processing an accepted event. Provider redelivery sends it through receipt again. A historical replay may use a different mechanism or payload representation depending on the provider.

Explain which paths preserve identity and which checks apply. Do not assume every replay tool behaves like the original delivery.

Keep signature verification and source authenticity appropriate to the supported path rather than disabling validation to make a replay succeed.

## Assign effect recovery

For local projections, replay may be straightforward under the application's duplicate and ordering rules. For external actions, require outcome inspection and the relevant operation recovery procedure.

Name the owner who can resolve an uncertain effect or approve a business correction. A generic replay button should not bypass that decision.

Record each recovery action and its result against the original event or operation.

## Practise a safe failed event

Have the receiving team locate a synthetic event that was accepted but not completed, fix its controlled failure and resume it. Then give them an uncertain external outcome and verify that they investigate before repeating the action.

Use the exercise to improve missing links between provider delivery history, inbox state and business evidence.

The handover is complete when another operator can account for an event and choose a safe recovery path. Knowing how to resend a payload is only a small part of that responsibility.

## Sources

- [GitHub: webhook best practices](https://docs.github.com/en/webhooks/using-webhooks/best-practices-for-using-webhooks)