fix: alerting rule

Author: codekow

components/operators/gpu-operator-certified/instance/components/gpu-monitoring/gpu-monitoring.yaml

@@ -14,8 +14,13 @@ spec:
         runbook_url: https://github.com/redhat-na-ssa/demo-ai-gitops-catalog/tree/main/components/operators/gpu-operator-certified/instance/components/monitoring-dashboard
         summary: Cloud costs may increase by requesting specialized resources.
       expr: |
-        sum (kube_pod_resource_request{resource="nvidia.com/gpu"} >= 1 ) > 0
+        # sum (kube_pod_resource_request{resource="nvidia.com/gpu"} >= 1 ) > 0
+        sum by (namespace, pod) (
+            kube_pod_resource_request{resource="nvidia.com/gpu"}
+          or
+            kube_pod_resource_limit{resource="nvidia.com/gpu"}
+        ) > 0
         # sum by (namespace, pod,resource) (kube_pod_resource_request{resource="nvidia.com/gpu"} >= 1) > 0
-      # for: 1m
+        # for: 1m
       labels:
         severity: info