Monitoring missing for metric output

There have been occasions in the past where either some or all metrics were missing from the output of prometheus-openstack-exporter, resolved by simply restarting the service.

We need to have some monitoring, either via NRPE plugin, or perhaps more reasonably by a prometheus rules export (using the https://github.com/CanonicalBootStack/interface-prometheus-rules interface), to ensure that a handful of common metrics are available. This would alert operators when the exporter is not running correctly and so intervention can take place to reduce or avoid gaps in metrics.