Merge pull request #225 from sysdiglabs/phpfpm

fix php-fpm alerts

Author: eckelon

resources/php-fpm/alerts.yaml

@@ -9,83 +9,46 @@ configurations:
 - kind: Prometheus
   data: |-
     groups:
-      - name: "Php-Fpm"
-        groupId: "php-fpm"
-        description: "Php-Fpm alerts"
-        integrationType: "php-fpm"
-        scopeVariables:
-          - variable: cluster
-            label: kubernetes.cluster.name
-            operator: in
-          - variable: namespace
-            label: kubernetes.namespace.name
-            operator: in
-          - variable: workload
-            label: kubernetes.workload.name
-            operator: in
+      - name: Php-Fpm
         rules:
-          - alert: "[Php-Fpm] Percentage of instances low"
-            alertId: "PercentageOfInstancesLow"
-            description: "Most of the instances are down"
+          - alert: '[Php-Fpm] Percentage of instances low'
             expr: |
-              sum by (kube_workload_name,kube_namespace_name,kube_cluster_name)(phpfpm_up{kube_cluster_name=~$cluster,kube_namespace_name=~$namespace,kube_workload_name=~$workload})/sum by (kube_workload_name,kube_namespace_name,kube_cluster_name)(kube_workload_status_desired{kube_cluster_name=~$cluster,kube_namespace_name=~$namespace,kube_workload_name=~$workload}) < 0.75
+              sum (phpfpm_up==1) / sum (phpfpm_up) < 0.75
             for: 5m
             labels:
-              severity: high
+              severity: critical
             annotations:
-              summary: |
-                [{{$labels.kube_cluster_name}} > {{$labels.kube_namespace_name}} > {{$labels.kube_workload_name}}] {{__alert_name__}} is {{__alert_status__}}
-              description: |
-                Most of the instances are down
-          - alert: "[Php-Fpm] Recently reboot"
-            alertId: "RecentlyReeboot"
-            description: "Instances have been recently reboot"
+              description: Less than 75% of instances are down
+          - alert: '[Php-Fpm] Recently reboot'
             expr: |
-              (count by (kube_cluster_name, kube_namespace_name,kube_pod_name)(phpfpm_start_since{kube_cluster_name=~$cluster,kube_namespace_name=~$namespace,kube_workload_name=~$workload} < 1800) or vector (0))/sum by (kube_cluster_name, kube_namespace_name,kube_pod_name)(phpfpm_up{kube_cluster_name=~$cluster,kube_namespace_name=~$namespace,kube_workload_name=~$workload}) > 0.4
+              (count (phpfpm_start_since < 1800) or vector (0))/sum (phpfpm_up) > 0.4
             for: 5m
             labels:
-              severity: high
+              severity: info
             annotations:
-              summary: |
-                [{{$labels.kube_cluster_name}} > {{$labels.kube_namespace_name}} > {{$labels.kube_workload_name}}] {{__alert_name__}} is {{__alert_status__}}
-              description: |
-                Instances have been recently reboot
-          - alert: "[Php-Fpm] Limit of child proccess exceeded"
-            alertId: "LimitOfChildsExceeded"
-            description: "Number of childs process have been exceeded"
+              description: Instances have been recently reboot
+          - alert: '[Php-Fpm] Limit of child proccess exceeded'
             expr: |
-              sum by (kube_cluster_name, kube_namespace_name,kube_pod_name) (rate (phpfpm_max_children_reached{kube_cluster_name=~$cluster,kube_namespace_name=~$namespace,kube_workload_name=~$workload}[5m])) > 0
+              sum (rate (phpfpm_max_children_reached[5m])) > 0
             for: 5m
             labels:
-              severity: high
+              severity: critical
             annotations:
-              summary: |
-                [{{$labels.kube_cluster_name}} > {{$labels.kube_namespace_name}} > {{$labels.kube_workload_name}}] {{__alert_name__}} is {{__alert_status__}}
-              description: |
-                Number of childs process have been exceeded
-          - alert: "[Php-Fpm] Reaching limit of queue process"
-            alertId: "ReachedQueueLimit"
-            description: "Buffer of queue requests reaching its limit"
+              description: Number of childs process have been exceeded
+          - alert: '[Php-Fpm] Reaching limit of queue process'
             expr: |
-              (phpfpm_listen_queue{kube_cluster_name=~$cluster,kube_namespace_name=~$namespace,kube_workload_name=~$workload} / phpfpm_listen_queue_length{kube_cluster_name=~$cluster,kube_namespace_name=~$namespace,kube_workload_name=~$workload}) > 0.8
+              (phpfpm_listen_queue / phpfpm_listen_queue_length) > 0.8
             for: 5m
             labels:
-              severity: medium
-            annotations:
-              summary: |
-                [{{$labels.kube_cluster_name}} > {{$labels.kube_namespace_name}} > {{$labels.kube_workload_name}}] {{__alert_name__}} is {{__alert_status__}}
-              description: |
-                Buffer of queue requests reaching its limit
-          - alert: "[Php-Fpm] Requests processed reached timeout limit"
-            alertId: "TimeoutLimitRequestReached"
-            description: "Timeout limit reached by some of the requests"
-            expr: |
-              rate (phpfpm_slow_requests{kube_cluster_name=~$cluster,kube_namespace_name=~$namespace,kube_workload_name=~$workload}[5m]) / rate (phpfpm_process_requests{kube_cluster_name=~$cluster,kube_namespace_name=~$namespace,kube_workload_name=~$workload}[5m]) > 0.75
-            for: 5m
-            labels:
-              severity: medium
-            annotations:
-              summary: |
-                [{{$labels.kube_cluster_name}} > {{$labels.kube_namespace_name}} > {{$labels.kube_workload_name}}] {{__alert_name__}} is {{__alert_status__}}
-              description: |
-                Timeout limit reached by some of the requests
+                severity: warning
+              annotations:
+                description: Buffer of queue requests reaching its limit
+            - alert: '[Php-Fpm] Too slow requests processing'
+              expr: |
+                rate (phpfpm_slow_requests[5m]) / rate (phpfpm_process_requests[5m]) > 0.1
+              for: 5m
+              labels:
+                severity: warning
+              annotations:
+                description: Requests have taking too much time to be processed
+