PG HA docs are light on example of usual failover
https://docs.gitlab.com/ee/administration/high_availability/alpha_database.html#failover-procedure
I'm not sure what normal failover looks like, so a little more detail of "look at this log, see this election" would be very useful
Context https://gitlab.com/gitlab-org/build/team-tasks/issues/14#note_47036464