I have had lots of success using streaming replication for potential failover as well as read only hot stand by to assist with my master db.
Additionally we use pg backrest for PITR in the event that we need to roll back for some unforeseen reason (hacking, corruption, etc).