In this recipe, we are creating a materialized view in Elasticsearch, and we want to allow users to search against the data in the region that they are closest to. However, Elasticsearch does not support regional replication. As we discussed in the Implementing regional replication with DynamoDB recipe, we do not want to replicate the event stream because that solution is complex and too difficult to reason about. Instead, as shown in the following diagram, we will place an S3 bucket in front of Elasticsearch in each region and leverage S3 triggers to update Elasticsearch and to implement a round-robin replication scheme:
