Attempt to make juniper alarms more robust

[chef.git] / cookbooks / prometheus / templates / default / alert_rules.yml.erb

diff --git a/cookbooks/prometheus/templates/default/alert_rules.yml.erb b/cookbooks/prometheus/templates/default/alert_rules.yml.erb
index 417641cf0b84943247359f2b6c214df128b52282..df1329147e5b6f585fdaa555c8565e771a31c645 100644 (file)
--- a/cookbooks/prometheus/templates/default/alert_rules.yml.erb
+++ b/cookbooks/prometheus/templates/default/alert_rules.yml.erb
@@ -217,12 +217,12 @@ groups:
   - name: juniper
     rules:
       - alert: juniper fan alarm
-        expr: jnxOperatingState{jnxOperatingContentsIndex="4",jnxOperatingState!="running"} > 0
+        expr: sum_over_time(jnxOperatingState{jnxOperatingContentsIndex="4",jnxOperatingState!="running"}[5m]) > 0
         for: 5m
         labels:
           alertgroup: "{{ $labels.site }}"
       - alert: juniper power alarm
-        expr: jnxOperatingState{jnxOperatingContentsIndex="2",jnxOperatingState!="running"} > 0
+        expr: sum_over_time(jnxOperatingState{jnxOperatingContentsIndex="2",jnxOperatingState!="running"}[5m]) > 0
         for: 5m
         labels:
           alertgroup: "{{ $labels.site }}"