The main building block in Akka Management is the Management HTTP server.
This HTTP server runs on a dedicated port, so you can separate the application traffic from the management traffic. It exposes, by default, some read-only endpoints where you can verify the status of the cluster or the health of the node.
To form a cluster between the different nodes of your application you can use hardcoded IP and port numbers. But that solution is very hard to maintain and goes against the mobility property of a reactive system.
Instead, Akka Management includes support to Bootstrap a cluster. Akka Cluster Bootstrap uses Akka Discovery (a service discovery API) to locate peers. This way, when deploying an application, each node only needs to register itself on the central registry and use the service discovery to find the peers.
The Akka Management Kubernetes Lease is an implementation of the Akka Coordination Lease which uses Kubernetes-backed mechanisms to increase the consistency of some Akka features like Split Brain Resolver , Akka Cluster Singleton , and Akka ClusterSharding .