how does Ambari manage components such as Apache Hadoop, Hive, and Spark?

Ambari now uses Apache Bigtop: https://bigtop.apache.org/

Bigtop is similar to HDP.

Can Ambari directly manage existing Hadoop clusters?

How do I get Ambari to manage and monitor my open source cluster? I already have data on my current Hadoop cluster and don't want to rebuild a new cluster.

Ambari can do this, but it's not an easy process. Much easier to deploy a Bigtop cluster from scratch using Ambari.

Using Ambari on top of an existing cluster requires creating Ambari Blueprints to post cluster info and configurations to Ambari Server. Some details here: https://www.adaltas.com/en/2018/01/17/ambari-how-to-blueprint/