feat(events): own the VM lifecycle, update-status, and restore-completion flows (5/7)

[marceloneppel]

Issue

The VM charm's lifecycle orchestration (install/start/leader-elected/config-changed bootstrap, update-status, restore completion, teardown, rolling restart, actions) still lives charm-side; the lib's existing skeletons predate the current managers.

Solution

• events/postgresql.py: faithful rewrite of the VM flows — leader-elected (password ensure incl. configured system-user secrets, members bootstrap, internal CA, endpoints), start (gates, member IP, storage layout, port, raft observer, bootstrap split), _start_primary (bootstrap + _setup_users + flags + extensions), _start_replica, the update-status family (+ _was_restore_successful, _handle_processes_failures), teardown (storage-detaching/remove), the rolling-ops _restart callback, and the get-primary/promote/secret handlers.
• Composition: the abstract builds s3_client/BackupManager/RestoreManager/ObserverManager and passes every bridge; CannotConnectError; the PRIMARY_NOT_REACHABLE_MESSAGE literal.
• Tests: tests/unit/test_vm_lifecycle_events.py (password ensure, update-status gates, teardown guards).

Checklist

• I have added or updated any relevant documentation.
• I have cleaned any remaining cloud resources from my accounts.