Can OpenSearch data-prepper be deployed in k8s deployment with multiple replicas without causing duplication in target / sink system?

I am planning to send logs to OpenSearch cluster on AWS using data-prepper log ingestion pipeline.