PG fully consistent automatic failovers are also pretty hard, at least they were 3-4 years ago when we did such a setup (initially it seemed easy, but tests sh

PG fully consistent automatic failovers are also pretty hard, at least they were 3-4 years ago when we did such a setup (initially it seemed easy, but tests showed its limits, and getting it reliable was pretty hard). Now I heard that there are Kubernetes based solution which are said to automate it pretty well, but have not tried those yet.