domain/monitor/alertmanager/datas.tf

locals {
  authentik_url = "http://authentik.${var.domain}-auth.svc"
  authentik_token = data.kubernetes_secret_v1.authentik.data["AUTHENTIK_BOOTSTRAP_TOKEN"]
  common_labels = {
    "vynil.solidite.fr/owner-name" = var.instance
    "vynil.solidite.fr/owner-namespace" = var.namespace
    "vynil.solidite.fr/owner-category" = var.category
    "vynil.solidite.fr/owner-component" = var.component
    "app.kubernetes.io/managed-by" = "vynil"
    "app.kubernetes.io/instance" = var.instance
  }
  rb-patch = <<-EOF
    - op: replace
      path: /subjects/0/namespace
      value: "${var.namespace}"
    EOF
}

data "kubernetes_secret_v1" "authentik" {
  metadata {
    name = "authentik"
    namespace = "${var.domain}-auth"
  }
}

data "kustomization_overlay" "data" {
  common_labels = local.common_labels
  namespace = var.namespace
  resources = [for file in fileset(path.module, "*.yaml"): file if file != "index.yaml" && length(regexall("ClusterRole",file))<1]
  patches {
    target {
      kind = "Alertmanager"
      name = "alertmanager-kube-promethe-alertmanager"
    }
    patch = <<-EOF
      apiVersion: monitoring.coreos.com/v1
      kind: Alertmanager
      metadata:
        name: alertmanager-kube-promethe-alertmanager
      spec:
        image: "${var.images.alertmanager.registry}/${var.images.alertmanager.repository}:${var.images.alertmanager.tag}"
        version: ${var.images.alertmanager.tag}
        externalUrl: http://${var.instance}-${var.component}.${var.namespace}:9093
        replicas: ${var.replicas}
        listenLocal: ${var.listenLocal}
        logLevel:  "${var.logLevel}"
        retention: "${var.retention}"
    EOF
  }
  patches {
    target {
      kind = "ConfigMap"
      name = "alertmanager-kube-grafana-datasource"
    }
    patch = <<-EOF
      apiVersion: v1
      kind: ConfigMap
      metadata:
        name: alertmanager-operated
      data:
        datasource.yaml: |-
          apiVersion: 1
          datasources:
            - name: Alertmanager
              type: alertmanager
              uid: alertmanager
              url: http://${var.instance}-${var.component}.${var.namespace}:9093/
              access: proxy
              jsonData:
                handleGrafanaManagedAlerts: false
                implementation: prometheus
    EOF
  }
  patches {
    target {
      kind = "ServiceMonitor"
      name = "alertmanager-kube-promethe-alertmanager"
    }
    patch = <<-EOF
    - op: replace
      path: /spec/namespaceSelector/matchNames/0
      value: "${var.namespace}"
    EOF
  }
  patches {
    target {
      kind = "PrometheusRule"
      name = "alertmanager-kube-promethe-alertmanager.rules"
    }
    patch = <<-EOF
apiVersion: monitoring.coreos.com/v1
kind: PrometheusRule
metadata:
  name: alertmanager-kube-promethe-alertmanager.rules
spec:
  groups:
  - name: alertmanager.rules
    rules:
    - alert: AlertmanagerFailedReload
      expr: |-
        # Without max_over_time, failed scrapes could create false negatives, see
        # https://www.robustperception.io/alerting-on-gauges-in-prometheus-2-0 for details.
        max_over_time(alertmanager_config_last_reload_successful{job="alertmanager-kube-promethe-alertmanager",namespace="${var.namespace}"}[5m]) == 0
    - alert: AlertmanagerMembersInconsistent
      expr: |-
        # Without max_over_time, failed scrapes could create false negatives, see
        # https://www.robustperception.io/alerting-on-gauges-in-prometheus-2-0 for details.
          max_over_time(alertmanager_cluster_members{job="alertmanager-kube-promethe-alertmanager",namespace="${var.namespace}"}[5m])
        < on (namespace,service,cluster) group_left
          count by (namespace,service,cluster) (max_over_time(alertmanager_cluster_members{job="alertmanager-kube-promethe-alertmanager",namespace="${var.namespace}"}[5m]))
    - alert: AlertmanagerFailedToSendAlerts
      expr: |-
        (
          rate(alertmanager_notifications_failed_total{job="alertmanager-kube-promethe-alertmanager",namespace="${var.namespace}"}[5m])
        /
          ignoring (reason) group_left rate(alertmanager_notifications_total{job="alertmanager-kube-promethe-alertmanager",namespace="${var.namespace}"}[5m])
        )
        > 0.01
    - alert: AlertmanagerClusterFailedToSendAlerts
      expr: |-
        min by (namespace,service, integration) (
          rate(alertmanager_notifications_failed_total{job="alertmanager-kube-promethe-alertmanager",namespace="${var.namespace}", integration=~`.*`}[5m])
        /
          ignoring (reason) group_left rate(alertmanager_notifications_total{job="alertmanager-kube-promethe-alertmanager",namespace="${var.namespace}", integration=~`.*`}[5m])
        )
        > 0.01
    - alert: AlertmanagerClusterFailedToSendAlerts
      expr: |-
        min by (namespace,service, integration) (
          rate(alertmanager_notifications_failed_total{job="alertmanager-kube-promethe-alertmanager",namespace="${var.namespace}", integration!~`.*`}[5m])
        /
          ignoring (reason) group_left rate(alertmanager_notifications_total{job="alertmanager-kube-promethe-alertmanager",namespace="${var.namespace}", integration!~`.*`}[5m])
        )
        > 0.01
    - alert: AlertmanagerConfigInconsistent
      expr: |-
        count by (namespace,service,cluster) (
          count_values by (namespace,service,cluster) ("config_hash", alertmanager_config_hash{job="alertmanager-kube-promethe-alertmanager",namespace="${var.namespace}"})
        )
        != 1
    - alert: AlertmanagerClusterDown
      expr: |-
        (
          count by (namespace,service,cluster) (
            avg_over_time(up{job="alertmanager-kube-promethe-alertmanager",namespace="${var.namespace}"}[5m]) < 0.5
          )
        /
          count by (namespace,service,cluster) (
            up{job="alertmanager-kube-promethe-alertmanager",namespace="${var.namespace}"}
          )
        )
        >= 0.5
    - alert: AlertmanagerClusterCrashlooping
      expr: |-
        (
          count by (namespace,service,cluster) (
            changes(process_start_time_seconds{job="alertmanager-kube-promethe-alertmanager",namespace="${var.namespace}"}[10m]) > 4
          )
        /
          count by (namespace,service,cluster) (
            up{job="alertmanager-kube-promethe-alertmanager",namespace="${var.namespace}"}
          )
        )
        >= 0.5
    EOF
  }
}