A standard SQS queue feeds workers that occasionally encounter malformed messages. These messages fail repeatedly and consume processing capacity. Operations must inspect them separately after several failed deliveries. Which configuration is most appropriate?

1. Have workers permanently ignore every message after its first failure.
2. Configure a dead-letter queue and a redrive policy with an appropriate maxReceiveCount.
3. Shorten all message retention to one minute so malformed and valid messages disappear quickly.
4. Enable long polling and assume it identifies invalid JSON messages automatically.