[Feature Request]: RequestResponseIO: Unable to configure retry Backoff parameters; defaults to unlimited retries

[jigarkb]

What would you like to happen?

Note: This is only applicable if we fix this: #37176

Problem: The RequestResponseIO transform currently hardcodes the retry backoff strategy to FluentBackoff.DEFAULT. This default strategy does not impose a limit on the number of retries or the total duration, effectively leading to unlimited retries for retryable errors (UserCodeRemoteSystemException, UserCodeTimeoutException).

Limitation: Users currently cannot override this behavior because the method responsible for configuring the backoff strategy is package-private

Impact: In production environments, unlimited retries are dangerous as they can cause pipelines to get stuck indefinitely on persistent errors or create backpressure issues. Users need the ability to define bounds

Location: https://github.com/apache/beam/blob/v2.67.0/sdks/java/io/rrio/src/main/java/org/apache/beam/io/requestresponse/RequestResponseIO.java#L225-L235

Current Code:

RequestResponseIO<RequestT, ResponseT> withBackOffSupplier(SerializableSupplier<BackOff> value) {
    return new RequestResponseIO<>(
        rrioConfiguration, callConfiguration.toBuilder().setBackOffSupplier(value).build());
  }

Suggested Fix:

public RequestResponseIO<RequestT, ResponseT> withBackOffSupplier(SerializableSupplier<BackOff> value) {
        return new RequestResponseIO<>(
                rrioConfiguration, callConfiguration.toBuilder().setBackOffSupplier(value).build());
    }

Issue Priority

Priority: 2 (default / most feature requests should be filed as P2)

Issue Components

• Component: Python SDK
• Component: Java SDK
• Component: Go SDK
• Component: Typescript SDK
• Component: IO connector
• Component: Beam YAML
• Component: Beam examples
• Component: Beam playground
• Component: Beam katas
• Component: Website
• Component: Infrastructure
• Component: Spark Runner
• Component: Flink Runner
• Component: Samza Runner
• Component: Twister2 Runner
• Component: Hazelcast Jet Runner
• Component: Google Cloud Dataflow Runner