add ECC alert (closes #19)

prometheus/alerts.yaml

@@ -81,6 +81,20 @@ spec:
         annotations:
           summary: "SMARTcheck not running"
           description: 'The last SMARTcheck on node {{ $labels.node }} was more than 3h ago. Plox fix.'
+      - alert: "ECC Memory errors"
+        expr: (node_edac_correctable_errors_total) > 100
+        labels:
+          severity: warning
+        annotations:
+          summary: "ECC errors on {{ $labels.node }}"
+          description: 'The node {{ $labels.node }} accumulated {{ $value }} correctable errors.'
+      - alert: "ECC Memory uncorrectable errors"
+        expr: (node_edac_uncorrectable_errors_total) > 0
+        labels:
+          severity: critical
+        annotations:
+          summary: "ECC errors on {{ $labels.node }}"
+          description: 'The node {{ $labels.node }} accumulated {{ $value }} uncorrectable errors.'
     - name: etcdbackup
       rules:
       - alert: "etcdbackup too old"
@@ -106,4 +120,5 @@ spec:
           severity: warning
         annotations:
           summary: "The Pod {{ $labels.pod }} is {{ $labels.reason }}"
-          description: "The Pod {{ $labels.pod }} is in the state {{ $labels.reason }} for more than 5m. The Pod is in namespace {{ $labels.namespace }} and on node {{ $labels.node }}."
+          description: "The Pod {{ $labels.pod }} is in the state {{ $labels.reason }} for more than 5m. The Pod is in namespace {{ $labels.namespace }} and on node {{ $labels.node }}."
+