We wonder about the high availability aspect of this solution. After all the master node can be down while with the conventional method, high availability is built in as we connect to a set of indexers.
If you're worried about that mechanism failing, you could also look at using DNS round robin by setting up a DNS record that points to all your indexers and then put that DNS name in your outputs.conf on your forwarders. Then you just need to make sure whenever you change the indexer cluster (e.g. add a node), to also update the DNS record.