increasing HA

Hello,

I am looking after some advice about solution allowing to increase High
availability?

Here is a bit of context :

I have an Master-Slave architecture
- 1 master
- 2 asynchronous slaves using replication slot
- backup is made with barman using replication slot
- Wal archiving is done towards barman server

I think 3 axes could be improved ( I am totaly novice with these):

- using of a proxy
I found HAproxy.
Could you advice any others solutions to explore or share your experience?

- using an automatick failover
I found PAF
Could you advice any others solutions to explore or share your experience?

- using a tool for fencing a failing node
Ideally, I aimagine to disable network traffic in I/O to prevent client
connecting and exchange between backup server failling server + on
postgesql server disable automatic restart of the service.
Could you share you experience about it?

- Maybe an other axe to explore ?

Thank you

Thomas