Merge pull request #634 from lilic/adjust-prom-operator-alerts

jsonnet/prometheus-operator.libsonnet: Adjust alerts range

jsonnet/kube-prometheus/alerts/prometheus-operator.libsonnet

@@ -7,7 +7,7 @@
           {
             alert: 'PrometheusOperatorListErrors',
             expr: |||
-              (sum by (controller,namespace) (rate(prometheus_operator_list_operations_failed_total{%(prometheusOperatorSelector)s}[1h])) / sum by (controller,namespace) (rate(prometheus_operator_list_operations_total{%(prometheusOperatorSelector)s}[1h]))) > 0.4
+              (sum by (controller,namespace) (rate(prometheus_operator_list_operations_failed_total{%(prometheusOperatorSelector)s}[10m])) / sum by (controller,namespace) (rate(prometheus_operator_list_operations_total{%(prometheusOperatorSelector)s}[10m]))) > 0.4
             ||| % $._config,
             labels: {
               severity: 'warning',
@@ -20,7 +20,7 @@
           {
             alert: 'PrometheusOperatorWatchErrors',
             expr: |||
-              (sum by (controller,namespace) (rate(prometheus_operator_watch_operations_failed_total{%(prometheusOperatorSelector)s}[1h])) / sum by (controller,namespace) (rate(prometheus_operator_watch_operations_total{%(prometheusOperatorSelector)s}[1h]))) > 0.4
+              (sum by (controller,namespace) (rate(prometheus_operator_watch_operations_failed_total{%(prometheusOperatorSelector)s}[10m])) / sum by (controller,namespace) (rate(prometheus_operator_watch_operations_total{%(prometheusOperatorSelector)s}[10m]))) > 0.4
             ||| % $._config,
             labels: {
               severity: 'warning',

manifests/prometheus-rules.yaml

@@ -1875,7 +1875,7 @@ spec:
         message: Errors while performing List operations in controller {{$labels.controller}}
           in {{$labels.namespace}} namespace.
       expr: |
-        (sum by (controller,namespace) (rate(prometheus_operator_list_operations_failed_total{job="prometheus-operator",namespace="monitoring"}[1h])) / sum by (controller,namespace) (rate(prometheus_operator_list_operations_total{job="prometheus-operator",namespace="monitoring"}[1h]))) > 0.4
+        (sum by (controller,namespace) (rate(prometheus_operator_list_operations_failed_total{job="prometheus-operator",namespace="monitoring"}[10m])) / sum by (controller,namespace) (rate(prometheus_operator_list_operations_total{job="prometheus-operator",namespace="monitoring"}[10m]))) > 0.4
       for: 15m
       labels:
         severity: warning
@@ -1884,7 +1884,7 @@ spec:
         message: Errors while performing Watch operations in controller {{$labels.controller}}
           in {{$labels.namespace}} namespace.
       expr: |
-        (sum by (controller,namespace) (rate(prometheus_operator_watch_operations_failed_total{job="prometheus-operator",namespace="monitoring"}[1h])) / sum by (controller,namespace) (rate(prometheus_operator_watch_operations_total{job="prometheus-operator",namespace="monitoring"}[1h]))) > 0.4
+        (sum by (controller,namespace) (rate(prometheus_operator_watch_operations_failed_total{job="prometheus-operator",namespace="monitoring"}[10m])) / sum by (controller,namespace) (rate(prometheus_operator_watch_operations_total{job="prometheus-operator",namespace="monitoring"}[10m]))) > 0.4
       for: 15m
       labels:
         severity: warning