I am planning to develop REST & reactive data processing units using vertx library/framework. I have a Kubernetes cluster where other services were deployed, and plan to deploy the new vertx based services too into this. Both relational and nosql databases are setup outside the kube cluster.

When it comes to vertx on Kubernetes, are there things I should consider when designing REST services based on vertx that reads and writes to relational database, or vertx based components that publish to or process messages from message brokers (say, ActiveMq or Rabbitmq or Kafka), or vertx file directory watcher to process newly created files?

Do I still have to consider Hazelcast or Zookeeper to run the services as cluster? How about the statelessness of the pods? The scalability?