Comment 1 for bug 1456797

Dolph, old revocation events already seem to be deleted as part of the list_events operation. See:

1) the call to _prune_expired_events() from list_events() in keystone/contrib/revoke/backends/sql.py
2) the call to _prune_expired_events_and_get() from list_events() in keystone/contrib/revoke/backends/kvs.py

It seems that list_events() is a high-traffic call (via the "GET /OS-REVOKE/events" API) that would therefore ensure that the event count in the database is not too large. Should we still augment the keystone-manage command line tool with the "revocation_event_flush" command (similar to "token_flush" command)? If so, what would be the reasons?