Route events using AWS EventBridge rules with automatic retries and a dead-letter queue for failed targets

Steps

1. Create a custom event bus or use the default bus; publish events via PutEvents API with Source, DetailType, and Detail (JSON object) fields — up to 10 events per PutEvents call
2. Create an EventBridge rule with an event pattern (JSON matching filter) or schedule expression targeting the event bus; specify one or more targets (Lambda, SQS, Step Functions, etc.)
3. Configure retry policy on each target: set MaximumRetryAttempts (0–185) and MaximumEventAgeInSeconds (60–86400); EventBridge retries with exponential backoff for failed target invocations
4. Attach a dead-letter queue (SQS) to each target's DeadLetterConfig so events that exhaust retries are captured; grant EventBridge the sqs:SendMessage permission on the DLQ via a queue resource policy
5. For Lambda targets, EventBridge invokes asynchronously — failures trigger the target's own retry/DLQ mechanism in addition to EventBridge retries; set the Lambda destination or function-level DLQ to avoid double-handling
6. Use event archive and replay (EventBridge Archive) on the event bus to store events for reprocessing; set retention to match your compliance or debugging window