landscape setup health checks

[Andrea Ieri]

A charmed installation of landscape does not really deploy health checks, therefore the only way for an operator to detect issues is logging into the UI.

For example (list not exhaustive), the following conditions are not reflected outside of the UI:
* landscape client cannot be authorized due to lack of available seats
* license overflow
* license expiration
* update_security_db.sh cannot run due to proxy issues

It would be useful if the above generated warnings in any of the following:
* landscape-client charm status
* nrpe check on client machine
* landscape-server charm status
* nrpe check on server machine

Given that we heavily rely on juju-lint and nagios to automate the validation and continuous monitoring of clouds, providing the above would be a big time saver.

[Drew Freiberger (afreiberger)]

Given the recent merge of nrpe-external-master relation (lp#1934816), we should revisit this RFE as implementation of additional health checks should now be trivial.

Here are some additional checks which would provide smoother daily operations:

* Pending computer registrations (critical if non-zero)
* Pending security updates (default warning, critical if over some configurable threshold)
* Clients not checking in for X days (critical)

Given this list and the list in original description, it seems it would be helpful if the entire Account Summary page action items/notifications could be exposed to the monitoring/alerting system.