Redfish resource health alerts when data not available.

[dashmage]

Currently, this is how the redfish metrics are created for the processor resource (it is similar for other redfish resources).
https://github.com/canonical/prometheus-hardware-exporter/blob/main/prometheus_hardware_exporter/collectors/redfish.py#L197

{
    "processor_id": processor["Id"],
    "model": processor["Model"] or "NA",
    "health": processor["Status"]["Health"] or "NA",
    "state": processor["Status"]["State"],
}

And this is how the alert rule corresponding to that metric looks like
https://github.com/canonical/hardware-observer-operator/blob/master/src/prometheus_alert_rules/redfish.yaml#L40

      - alert: RedfishProcessorHealthNotOk
        expr: redfish_processor_info{health != "OK"}
        for: 0m
        labels:
          severity: critical
        annotations:
          summary: Redfish processor health not OK. (instance {{ $labels.instance }})
          description: |
            Redfish processor health not OK.
              LABELS = {{ $labels }}

If the processor health is not available we assign it the value of "NA". This also triggers a critical alert (since it's not "OK") which might not be necessary just if the processor health status cannot be queried.

If needed, we could create a new alert with a lower severity if the resource health data is not available (with value "NA").

[dashmage]

#112 is related to this.