I do, but I take a very cautious approach. I run a custom image with PostgreSQL and Patroni on Kubernetes, no operator, each replica has it's own StatefulSet ti

I do, but I take a very cautious approach. I run a custom image with PostgreSQL and Patroni on Kubernetes, no operator, each replica has it's own StatefulSet tied to a specific node. There is very little automation, but it still better than running PostgreSQL outside of Kubernetes. I get the benefit of simplified monitoring, log handling, request routing, while still having very static resource assignments.