Add unit test to alertmanager.rules

helm/prometheus-rules/templates/alerting-rules/alertmanager.rules.yml

@@ -15,8 +15,11 @@ spec:
       annotations:
         description: '{{`AlertManager {{ $labels.integration }} notifications are failing.`}}'
         opsrecipe: alert-manager-notifications-failing/
-      expr: rate(alertmanager_notifications_failed_total{integration!="opsgenie"}[5m]) > 0
-      for: 10m
+      # Interval = 20m because currently AlertManager config set `group_interval=15m` that means that if a notification fails, it will be retried after 15m
+      # so the counter will stay flat during this time.
+      # Here, we decide to page after 3 successive failures, so we need to wait 3*15m = 45m before paging.
+      expr: rate(alertmanager_notifications_failed_total{integration!="opsgenie"}[20m]) > 0
+      for: 45m
       labels:
         area: empowerment
         severity: page
@@ -27,8 +30,9 @@ spec:
       annotations:
         description: '{{`AlertManager {{ $labels.integration }} notifications are failing.`}}'
         opsrecipe: alert-manager-notifications-failing/
-      expr: rate(alertmanager_notifications_failed_total{integration="opsgenie"}[5m]) > 0
-      for: 2m
+      # Here, we decide to notify after 2 successive failures (opsgenie notification), so we need to wait 2*15m = 30m before notifying.
+      expr: rate(alertmanager_notifications_failed_total{integration="opsgenie"}[20m]) > 0
+      for: 30m
       labels:
         area: empowerment
         severity: notify

test/conf/promtool_ignore

@@ -1,6 +1,5 @@
 templates/alerting-rules-test/capo.rules.test.yml
 templates/alerting-rules/alertmanager-dashboard.rules.yml
-templates/alerting-rules/alertmanager.rules.yml
 templates/alerting-rules/apiserver.management-cluster.rules.yml
 templates/alerting-rules/apiserver.workload-cluster.rules.yml
 templates/alerting-rules/argocd.rules.yml

test/tests/providers/global/alertmanager.rules.test.yml

@@ -0,0 +1,62 @@
+---
+rule_files:
+  - alertmanager.rules.yml
+
+tests:
+  - interval: 1m
+    input_series:
+      # after 1h, slack notification fails during 2h then works again (15m group_interval) => alert fires after 3 successive failures
+      - series: 'alertmanager_notifications_failed_total{integration="slack"}'
+        values: "0x60 1+0x15 2+0x15 3+0x15 4+0x15 5+0x15 6+0x15 7+0x15 8+0x15 8+0x120"
+      # after 1h, slack notification fails 2 times during 30mn than works again => alert must not fires
+      - series: 'alertmanager_notifications_failed_total{integration="webhook"}'
+        values: "0x60 1+0x15 2+0x15 2+0x15 2+0x15 2+0x15 2+0x15 2+0x15 2+0x15 2+0x120"
+    alert_rule_test:
+      - alertname:  AlertmanagerPageNotificationsFailing
+        eval_time: 10m
+      - alertname:  AlertmanagerPageNotificationsFailing
+        eval_time: 90m
+      - alertname:  AlertmanagerPageNotificationsFailing
+        eval_time: 95m
+      - alertname:  AlertmanagerPageNotificationsFailing
+        eval_time: 106m
+        exp_alerts:
+          - exp_labels:
+              area: empowerment
+              severity: page
+              team: atlas
+              topic: monitoring
+              integration: slack
+              cancel_if_outside_working_hours: "true"
+            exp_annotations:
+              description: "AlertManager slack notifications are failing."
+              opsrecipe: alert-manager-notifications-failing/
+      - alertname:  AlertmanagerPageNotificationsFailing
+        eval_time: 240m
+  - interval: 1m
+    input_series:
+      # after 1h, opsgenie notification fails during 45m then works again for 1h, finally fails 1 time (group_interval=15m)
+      # => alert fires after 2 successive failures only
+      - series: 'alertmanager_notifications_failed_total{integration="opsgenie"}'
+        values: "0x60 1+0x15 2+0x15 2+0x15 2+0x60 3+0x15 3+0x60"
+    alert_rule_test:
+      - alertname:  AlertmanagerNotifyNotificationsFailing
+        eval_time: 10m
+      - alertname:  AlertmanagerNotifyNotificationsFailing
+        eval_time: 75m
+      - alertname:  AlertmanagerNotifyNotificationsFailing
+        eval_time: 91m
+        exp_alerts:
+          - exp_labels:
+              area: empowerment
+              severity: notify
+              team: atlas
+              topic: monitoring
+              integration: opsgenie
+            exp_annotations:
+              description: "AlertManager opsgenie notifications are failing."
+              opsrecipe: alert-manager-notifications-failing/
+      - alertname:  AlertmanagerNotifyNotificationsFailing
+        eval_time: 180m
+      - alertname:  AlertmanagerNotifyNotificationsFailing
+        eval_time: 210m