Hi Jerome, well, this setup is quite stable, I configured this schema few time in production for the medium-load web project. In my opinion, stability is more appears to the networking and servers parameters, I mean the count of requests that you’ll able to serve depends on network speed + CPU/Mem capability + Disk subsystem speed (SSD disks in any failover RAID will be OK). One thing that I have changed after this article was written, the Heartbeat. Now I use Keepalived, it’s a more popular and stable/modern solution.

In my last tutorial, you can find a simple schema of Keepalived VRRP configuration:
https://medium.com/faun/failover-redis-like-cluster-from-two-masters-with-keydb-9ab8e806b66c

I hope this was helpful response :)