Quarkus Amazon Clients and resilience

[tor-jorgen-yara]

The Amazon SDKs implement automatic retry logic, but how should resilience be handled from the Quarkus Amazon Clients?
For the asynchronous DynamoDB client, the only things I can see that can be configured is:

• quarkus.dynamodb.async-client.xxx-timeout

Is the underlying functionality in the SDKs turned off, so that we can use smallrye-fault-tolerance for this, or how should this be handled correctly?

[gsmet]

I think it's just that it hasn't been implemented.

The Amazon Services extensions have been extracted here https://github.com/quarkiverse/quarkus-amazon-services for easier contribution so feel free to contribute there.

[tor-jorgen-yara]

Thanks for replying Guillaume. I had a look at the code and the default AWS retry policy is intact. I'll create a PR if we need to tweak the default policy...