Dead-letter queue
In Arabic قائمة الرسائل المتعثرة
A dead-letter queue is a holding place for messages between systems that still failed after their retries, kept so they can be inspected, replayed or cancelled rather than lost.
In plain words
When one system sends another a message and the other end is down or rejects it, the message has to go somewhere. Retrying forever blocks everything behind it; dropping it loses a customer or an order. The dead-letter queue is the middle path: keep it aside and let a person decide.
A message is retried a set number of times, often with growing gaps. If it still fails, it moves to the dead-letter queue with the reason. Once the cause is fixed, such as an address corrected or a missing code added, the messages are replayed, one by one or in bulk.
The common mistake is a dead-letter queue nobody watches. Failures pile up quietly while both systems look healthy, and by the time someone looks, hundreds of records are out of step. Show its size where the team will see it, and clear it as part of the routine.
- Messages sent
- 240
- Delivered first time
- 228
- Delivered on a retry
- 9
- Waiting in the dead-letter queue
- 3
In Axon Suite
How Axon Suite handles it
In Axon Suite a failed webhook call is retried, and what still fails is kept for you to replay or cancel in bulk. The health view shows what is waiting in the dead-letter queue.
Feature
Signed webhooks you can replay
Axon Suite posts each new customer to your own middleware through your field map, signed so you know it came from you. Failures are retried, kept for replay, and a kill switch stops sending while you fix the other end.
Open: Signed webhooks you can replay