This is an on-premise client cloud implementation using Pega 8.6.2 and PostgreSQL 11. The client has implemented a master-slave HA configuration with synchronous fsync replication to ensure 0 data loss. We have observed massive regular Commit delays of up to 27 seconds when the WAL file gets replicated and flushed to disk. There's no excessive commits in the application itself and it is kept to an absolute minimum. This is a REST API based transaction.

My question is whether anyone out there has successfully implemented PostgreSQL for High Availability possibly using third-party solutions. If so it would be much appreciated if you could provide as much details as possible.

Also if you have any tuning recommendations for a synchronous master-slave HA implementation as mentioned above it would also be much appreciated.