EventBridge:Retry policies and dead-letter queues

Retry policies

A retry policy specifies the method that is used to retry the events that failed to be sent. Event streams in EventBridge supports the backoff retry and exponential decay retry policies.

• Backoff retry: A failed event can be retried for up to three times. The interval between two consecutive retries is a random value that ranges from 10 to 20 seconds. This is the default retry policy.

• Exponential decay retry: A failed event can be retried for up to 176 times. The interval between two consecutive retries exponentially increases to a maximum of 512 seconds. The total amount of time that can be used for retries is one day. The specific intervals are 1 second, 2 seconds, 4 seconds, 8 seconds,16 seconds, 32 seconds, 64 seconds, 128 seconds, 256 seconds, and 512 seconds. The value 512 seconds is used for 167 retries.

Fault tolerance policies

A fault tolerance policy specifies the method that is used to handle errors. Event streams in EventBridge support the following fault tolerance policies:

• Fault tolerance allowed: allows fault tolerance. If an error occurs, event processing is not blocked. If an event fails to be sent after the specified maximum number of retries is reached, the event is delivered to the dead-letter queue or discarded.

• Fault tolerance prohibited: does not allow fault tolerance. If an error occurs in an event and the event fails to be sent after the specified maximum number of retries is reached, event processing is blocked. In this case, the status of the task is automatically changed to Ready.

Dead-letter queues