fixed alerts

k8s/core/prom-stack/grafana-alerting-configmap.yaml

@@ -260,7 +260,7 @@ data:
                   to: 0
                 datasourceUid: P76F38748CEC837F0
                 model:
-                  expr: 'node_load5 / on(instance) group_left count by(instance)(node_cpu_seconds_total{mode="idle"})'
+                  expr: 'node_load15 / on(instance) group_left count by(instance)(node_cpu_seconds_total{mode="idle"})'
                   refId: A
                   intervalMs: 1000
                   maxDataPoints: 43200
@@ -273,7 +273,7 @@ data:
                   conditions:
                     - evaluator:
                         params:
-                          - 0.8
+                          - 2
                         type: gt
                       operator:
                         type: and
@@ -283,16 +283,16 @@ data:
                     type: __expr__
                     uid: __expr__
                   expression: A
-                  reducer: max
+                  reducer: last
                   refId: B
                   type: reduce
             noDataState: NoData
             execErrState: Alerting
-            for: 5m
+            for: 15m
             annotations:
               node: '{{ $labels.instance }}'
               load_average: '{{ printf "%.2f" $values.A }}'
-              summary: 'Node load average is high relative to CPU count'
+              summary: 'Node load average is critically high relative to CPU count'
             labels:
               severity: warning