Hi
#Amplify AppSync - Integrate PostgreSQL
steep sail
bitter beacon
Aurora is the way to go. It’s not really a ‘server less’ like Lambda or Dynamo for example, but it’s a serverless RDS in a way that it starts with a minimal resource instance which is required to be always up, and scales the number of nodes up to the maximum specified buy you when it encountering more workload. This means that even if you don’t have any calls to Aurora, you will have to pay because it requires to always have a node with minimum resources up. There is no ‘pay per request’ feature here like you could have on Dynamo. In your case(I read your previous thread), it’s better to get use of @Searchable and OpenSearch. The costs will be less then having an RDS or Aurora cluster. For development environments you can use just the smallest t3.small instance, and for productive environments you can specify a more powerful one based on your needs.
Just compare the costs with an RDS or Aurora. It will be much less
I don’t think it’s straight forward to connect a RDS to the AppSync resolver within Amplify context but here are my thought on how I’ll try to implement it: 1. Have a separated Aurora Cluster out of Amplify context. 2 User ApiGateway as API and not AppSync. Do your logic in lambdas and consume the data via APIGateway. You will have to generate all the models by yourself.
Yes. Btw, did you took a look at this examples? https://docs.amplify.aws/cli/graphql/examples-and-solutions/#warehouse-management-system
Maybe you can solve everything you need by applying these concepts to your use case
The problem we have with the searchable annotation is described here: https://discord.com/channels/705853757799399426/1095618466800730193. If you are only one which works on the project, you’ll be fine. For our production environment, we have just one Aurora Cluster and one OpenSearch cluster. The price for Aurora > OpenSearch, but it can get reverted. It depends of the workloads and instance types
If you are in the free-tier it will have 750h free of usage, so basically free. The prices per hour can be found here: https://aws.amazon.com/opensearch-service/pricing/
If not, the price per hour for the smallest instance is 0,036*750=27 USD
Just for the first year
By default Amplify sets the smallest instance when you add @searchable annotation. You can override the resource, also within Amplify context. Just check the docs
It creates multiple indexes within the same OpenSearch domain. But if you have different or multiples environments, ex prod, test and so one, every environment will have a separated OpenSearch domain