During upgrade and reboot of cloudcontrol nodes today, quite a few alerts managed to fire despite the cookbook attempting to downtime things. Here's most of them:

• Failure in check of cinder snapshots: unable to read output
• NRPE failures:
• cloudinfra project instance distribution
  • Disk space
  • Puppet last run
  • dhclient process
  • Check for snapshots leaked by cinder backup agen
• Ensure NFS exports are maintained for new instances with NFS
• CRITICAL - Expecting active but unit nfs-exportd is activating