Since the backup cron is now a systemd timer to reduce unread email spam and improve monitoring (a serious alert by icinga needs to be tested still, but in theory it should catch a systemd timer gone bad that exits with non-zero). It seems like we should move other crons to that. The backup cron from the "secondary" file server role is the best example of one working in puppet.
A good example of where I think we'd benefit is the issues with emails coming out from the keystone jobs.
- profile::base::labs
- profile::wmcs::tenants::libraryupgrader
- profile::wmcs::monitoring
- profile::toolforge::clush::master
- role::prometheus::labs_project
- role::prometheus::tools
- role::labs::db::check_private_data
- role::labs::nfs::secondary
- role::openldap::labs
- dnsrecursor::labsaliaser
- labstore::fileserver::exports
- graphite::wmcs::archiver
- openstack::keystone::cleanup
- openstack::glance::image_sync
- openstack::wikitech::web
- openstack::designate::dns_floating_ip_updater
- openstack::puppet::master::instance_info_dumper