smallrye-reactive-messaging-kafka health check slower on RHBQ 2.13.5

Issue

• Moving from the Red Hat Build of Quarkus (RHBQ) 2.7 to 2.13.5, a generic reactive messaging kafka microservice such as described at Getting Started to SmallRye Reactive Messaging with Apache Kafka deployed on Openshift Container Platform (OCP) 4.10 using health-topic-verification-enabled=true and the default timeout of 1s is intermittently failing health checks.

• When the readiness is slow, the checks are duplicated:

$ curl localhost:8080/q/health/ready

{
    "status": "UP",
    "checks": [
        {
            "name": "Kafka Streams state health check",
            "status": "UP",
            "data": {
                "state": "RUNNING"
            }
        },
        {
            "name": "Kafka Streams state health check",
            "status": "UP",
            "data": {
                "state": "RUNNING"
            }
        },
        {
            "name": "Kafka Streams topics health check",
            "status": "UP",
            "data": {
                "available_topics": "skube-java-sdk-dev-samples.ping-events,skube-java-sdk-dev-samples.pong-events,skube-java-sdk-dev-samples.global-dlq,skube-java-sdk-dev-samples.skube-kstream-sample-dlq"
            }
        },
        {
            "name": "Kafka Streams topics health check",
            "status": "UP",
            "data": {
                "available_topics": "skube-java-sdk-dev-samples.ping-events,skube-java-sdk-dev-samples.pong-events,skube-java-sdk-dev-samples.global-dlq,skube-java-sdk-dev-samples.skube-kstream-sample-dlq"
            }
        }
    ]
}