Add alerting for critical node or hardware status

Depends on #7 (closed)

Once alerting is in place, it would be good to add alerts for critical events that aren't monitored. Prometheus Operator shipped with a fantastic set of alerts, but some metrics could be added. At least:

• temperature approaching maximum
• temperature exceeding maximum
• temperature approaching critical
• temperature exceeding critical

Could be added. Although that last one is likely to kill node_exporter, the first three would be very useful.