Automate cloud status prior to upgrade

[BjoernT]

Based in discussions with @galstrom21 I think it is beneficial to check the environment automatically on checks I usually do prior to scheduling upgrades:

• Are nova, neutron, cinder, heat services checking in (service-list and agent health commands)
• Is VXLAN used with neutron still in containers
• Hosts configured but not accessible inside the OSA inventory
• Custom cinder drivers used which are not included at upstream
• Hardware not on Ubuntu HCL
• Open MAAS/Monitoring alerts
• Customizations like forked OSA playbooks/code
• Customizations of Openstack Policies
• Last time OS patches had been installed and how many are missing
• Galera and RabbitMQ cluster checks are mandatory
• Mismatch of endpoint catalog and External/Internal_lb_vip_address
• Are Holland backups working (or is the control plane backed up by other means)
• Are multiple containers from the same type existing but not used
• Zombie processes
• iscsi errors on nova-compute consoles
• Did API service report recent errors
• Intrusive SSH options like limit sessions within sshd_config or authorized_keys
• git repos origin doesn't point to upstream
• git repo branch tracking misconfigured

[antonym]

#363 splits the preflight stuff so that leap and incremental can use them both and then we could build on top of that framework.
I'd imagine a lot of those are just automating the manual steps in the runbook docs.

There are also some healthcheck plays we're might be able to leverage for service checks so that we're not having to write stuff from scratch:

https://github.com/openstack/openstack-ansible/blob/stable/stein/playbooks/healthcheck-hosts.yml
https://github.com/openstack/openstack-ansible/blob/stable/stein/playbooks/healthcheck-infrastructure.yml
https://github.com/openstack/openstack-ansible/blob/stable/stein/playbooks/healthcheck-openstack.yml