I came across this bug [1] in triage today and I thought this was fixed already [2] but either something regressed or there is more to do here. I'm mostly just wondering, are operators already running any kind of script which purges old instance_faults table records before an instance is deleted and archived/purged? Because if so, that might be something we want to add as a nova-manage command. [1] https://bugs.launchpad.net/nova/+bug/1800755 [2] https://review.openstack.org/#/c/409943/ -- Thanks, Matt