Webhook Delivery Failures
Delivery failures are a normal part of webhook operations. Networks experience timeouts, services restart, and configurations change. How failures are handled determines the overall reliability of the system.
Failure Classification
Delivery failures are classified to determine the appropriate response:
| Failure Type | Examples | Handling |
|---|---|---|
| Transient | Network timeout, connection refused, rate limit | Automatic retry with backoff |
| Permanent | Invalid destination, authentication failure | Move to DLQ, no retry |
| Undetermined | Timeout without response | Retry, then DLQ if repeated |
Retry Policy
Retry behavior is configurable per destination:
- Maximum retry attempts
- Backoff interval and strategy (fixed, exponential)
- Retry conditions (which failure types trigger retry)
- Timeout per attempt
Dead-Letter Queue
When retries are exhausted, events move to the dead-letter queue. DLQ preserves:
- Original event payload and headers
- Delivery attempt history with timestamps
- Failure reason and classification
- Source and destination identifiers
Learn more about Dead Letter Queue
Recovery
From the DLQ, operators can:
- Inspect failure details and payload
- Replay events to original or alternate destinations
- Archive for compliance or audit
- Discard after review
Learn more about Replay and Recovery