Webhook Replay vs Retry
Retry and replay are complementary but serve different purposes in webhook delivery recovery.
Retry (Automatic)
Retry is an automatic response to delivery failure. When a delivery attempt fails with a transient error (timeout, connection refused), the system automatically retries according to the configured policy.
| Property | Retry |
|---|---|
| Trigger | Automatic on failure |
| Scope | Single event, same destination |
| Timing | Immediate with backoff |
| Control | Max attempts, backoff, conditions |
| User action | None (automatic) |
Replay (Manual or Automated)
Replay is a recovery action that resends events from the DLQ or delivery history. It is typically triggered by an operator after resolving the root cause of a failure.
| Property | Replay |
|---|---|
| Trigger | Manual or automated after resolution |
| Scope | Multiple events, multiple destinations |
| Timing | After failure diagnosis and resolution |
| Control | Event selector, destination, window |
| User action | Initiate replay, select events |
When to Use Each
| Scenario | Approach |
|---|---|
| Network blip (brief outage) | Retry handles automatically |
| Destination service restart | Retry handles with backoff |
| Configuration error | Replay after fixing config |
| Permanent destination change | Replay to new destination |
| Data loss from source | Replay from delivery history |