postgres-operator as highly available

[uk1988]

We are running postgres-operator in our small k8s clusters which are installed on-prem.
The cluster is set up as a 3 nodes cluster with all nodes running all k8s roles.
Per definition, postgres-operator has a single replica count. Is it possible to run postgres-operator replica count higher as 1?

With postgres-operator we deploy 2 DB one for keycloak and one for our application.
Both of the DB we do run as HA. However, postgres-operator we run only as a single pod since replica count is not a config parameter of the operator. What is the impact if the nodes where postgres-operator is running goes down for 2 hours?

[jfmatth]

does the operator have any bearing on a running cluster? I would think not?

[uk1988]

Yes in the end that is the question. If there is no bearing on a running cluster then there is no problem if it fails.