resilience4j: [QUESTION] Using circuit breaker and retry together

Resilience4j version: v1.4.0

Java version: 11

Hey team, I had a few questions regarding the use of CB and retry together:

1. Is there any documentation out there for how to get that set up? I’m confused as to how the retry fallbacks and circuit breaker fallbacks and state transitions work together.
2. My use case is to send some data to a Kafka cluster through a decorated Consumer and I’m using the decorateConsumer method from a CircuitBreaker but to execute it I don’t see any methods. Currently I’m just calling the accept method from a Consumer interface, is there any other way to run a decorated Consumer?
3. If I want to decorate a Consumer with both CircuitBreaker and Retry is there any way to do so without using the decorators dependency? I’m thinking of having just the r4j-circuitbreaker and r4j-retry dependencies and decorate the Consumer with one and save it in a variable and then decorate the new saved variable with the other. Thoughts on this?

EDIT:

1. For custom configuration for either decorators, if I use a custom config and only set a few of the options available, what happens for the other ones. For example if I use CircuitBreakerConfig object and build a custom config that only defines a failure threshold in half open state and closed state, what happens to all the other configuration properties that can be set such as number of calls permitted in half-open state, etc. Does it still use the default values for all the other properties and only overrides the ones I’ve set in the custom config object that is passed to the CircuitBreakerRegistry or CircuitBreaker directly?
2. Is there any retry functionality built into CircuitBreaker where if an error occurs then it stores the data that was supposed to be sent and then re-sent it. And if Retry is supposed to be used for that purpose, does Retry have a buffer where it stores the data that should’ve been sent and that it needs to retry sending? Or does the Retry functionality work on each call of a Consumer/Supplier/etc. at a time so it doesn’t care if 100 calls fail in a second because it’ll retry it with whatever logic is set up for each of those 100 calls separately in the same or new thread(s)?