CNINode resources' name should refect to node's instance ID

**What would you like to be enhanced**:
Currently the CNINode resources' name are node hostname which is likely reused by EC2. To be more resilient, using instance ID should be a better way to prevent collision.

**Why is the change needed and what use case will it solve**:
Hypothetically, if the resources were leaked, the name collision would cause breaking behavior on this controller.

Start by tracing how CNINode resource names are derived from the node hostname and where the node's EC2 instance ID is available. Verify the change against the controller's existing resource-handling behavior, and confirm that reused hostnames no longer produce a collision.

Written by the indexing model from the issue text.