Why do we want to use DynamoDB/S3 and integrate it with OpenSearch?
Why do we want to include other storage and duplicate the data between both of them.
Also I have a use case where I would need to do perform queries using different fields/columns.
Firstly I considered DynamoDB but I had to create new GSI for every new query. Creating so many GSIs is like creating tables itself and writing each time to all GSIs would increase latency and cost.
Secondly DocumentDB, again it supports ad-hoc queries but if the index is not predefined it would scan all the documents which i could still achieve using DynamoDb scan but that is not intended.
and creating the indexes in documentDB is also suggested to limit to 5 and I might have more than 5.
Next I explored OpenSearch but it shows everywhere to integrate with other primary database and i’m not sure why that needs to be done.