The restore logic lives in odoo/service/db.py (or openerp/service/db.py in older versions), in the restore_db function. That's the cleanest place to hook in your audit logging.

Option 1; Patch restore_db directly:

def restore_db(db, dump_file, copy=False):
    # ... existing restore logic ...
    
    # Add at the end, after restore completes:
    import logging, datetime
    _logger = logging.getLogger(__name__)
    _logger.info("DB RESTORE AUDIT: database=%s, backup=%s, timestamp=%s", 
                 db, dump_file, datetime.datetime.utcnow().isoformat())

Option 2; Write to a file (better for audit trail):

import datetime, os

audit_log = "/var/log/odoo/restore_audit.log"
with open(audit_log, "a") as f:
    f.write("{} | db={} | file={}\n".format(
        datetime.datetime.utcnow().isoformat(), db, dump_file
    ))

Option 3; PostgreSQL trigger (DB-level audit):

Create a table in a separate audit database or in postgres itself and insert from the restore script. This survives even if the restored DB overwrites application-level records.

Important caveat: since a restore replaces the database, any audit record written into the restored DB will be overwritten too. Always write the audit log to an external file, a separate DB, or the system logger; never into the database being restored.