Endpoints are not updated if the leader's patroni service is not running

[dragomirp]

Likely we will need to set the alternative endpoints parameter when getting the primary endpoint here:

postgresql-operator/src/charm.py

Line 390 in 7686dd0

primary = self._patroni.get_primary()

Since follower units can't update the endpoints and the leader cannot get the state of the cluster locally.

[syncronize-issues-to-jira]

Thank you for reporting your feedback to us!

The internal ticket has been created: https://warthogs.atlassian.net/browse/DPE-6874.

This message was autogenerated

[dragomirp]

Upon further investigation it seems to be that the cluster topology observer doesn't poll other members, so it will crash until service is recovered on the leader unit.

[dragomirp]

Should be fixed on latest 14/edge (revs. 589, 590)