Vert.x Event Bus and Hazelcast clustering discovery across environments

A Vert.x application using the event bus operates in a single-node local setup without explicit clustering configuration. The same application is intended to run as a multi-node deployment where the event bus should be clustered via Hazelcast.

Vert.x delegates cluster formation to Hazelcast and leaves the choice of discovery mechanism to the operator. In environments where multicast discovery is restricted or network ports are blocked, the default discovery behavior is uncertain and may lead to inconsistent cluster formation between development and production.

What discovery mechanism does Vert.x use by default for Hazelcast-based event bus clustering, and is it consistent between local and multi-node deployments? How is the Hazelcast discovery configuration exposed and required in Vert.x for consistent event bus formation when multicast is unavailable? Is a production default documented, or is the discovery choice left entirely to the operator?