DPE-7726: Start observer on non-Primary Patroni start to speedup re-join

The current PG16 logic relies on Juju update-status or on_topology_change
observer events, while in some cases we start Patroni without the Observer,
causing a long waiting story till the next update-status arrives.

Author: taurus-forever

src/charm.py

@@ -1927,6 +1927,7 @@ def _on_update_status(self, _) -> None:
 
         if not self._patroni.member_started and self._patroni.is_member_isolated:
             self._patroni.restart_patroni()
+            self._observer.start_observer()
             return
 
         # Update the sync-standby endpoint in the async replication data.
@@ -2056,8 +2057,9 @@ def _handle_processes_failures(self) -> bool:
                 logger.info("PostgreSQL data directory was not empty. Moved pg_wal")
                 return True
             try:
-                self._patroni.restart_patroni()
                 logger.info("restarted PostgreSQL because it was not running")
+                self._patroni.restart_patroni()
+                self._observer.start_observer()
                 return True
             except RetryError:
                 logger.error("failed to restart PostgreSQL after checking that it was not running")