ApsaraMQ for Kafka:Retry policies and dead-letter queues

Retry policies

A retry policy defines how a failed task is retried. Message Integration supports two retry policies: backoff retry and exponential decay retry.

• Backoff retry (Default): Retries a failed task up to three times. The retry interval is a random value between 10 seconds and 20 seconds.

• Exponential decay retry: Retries a failed task up to 176 times over a total duration of one day. The retry interval increases exponentially to a maximum of 512 seconds. The specific retry intervals are 1, 2, 4, 8, ..., 512 seconds.

Fault tolerance policies

A fault tolerance policy defines how errors are handled. Two types of fault tolerance policies are supported:

• Fault tolerance allowed (Recommended): Allows exceptions to occur without blocking task execution. If the number of retries exceeds the maximum limit, the message is delivered to a dead-letter queue or discarded based on your configuration.

• Fault tolerance prohibited: If an exception occurs and the number of retries exceeds the limit set in the retry policy, task execution is blocked. The task status then changes to Ready.

Dead-letter queues