retry topic
A topic that holds records whose processing failed so they can be tried again later without blocking the main topic.
In more detail
A chain of numbered retry topics models a retry schedule, with the dead letter queue as the last step. Each record carries an attempt counter so the router knows where to send it next.
Example
TIERS = ['orders', 'orders-retry-1', 'orders-retry-2', 'orders-dlq']Related terms
- dead letter queue — A separate queue where messages that keep failing are moved, so they stop blocking others and can be inspected and replayed later.
- attempt counter — A number stored in a record that counts how many times processing it has been tried.
- exponential backoff — Waiting progressively longer between retries, for example 1, 2, then 4 seconds, so a struggling service gets room to recover.
- transient failure — A failure caused by a temporary condition, such as a brief network glitch, that may succeed if simply tried again.
Taught in
- Retries With Backoff Topics — Apache Kafka & Event Streaming