Terraform：参考 YAML 列表

Question

我正在使用 Terraform 创建 Grafana 警报，并且创建了一个 YAML 配置，我试图在 Terraform 模块中引用该配置，但在引用嵌套列表时遇到了一些问题。

我的 YAML 配置“rules.yaml”：

dashboard_uid: "my-dashboard"
env:
  staging:
    folder_uid: "xxxxxxxxxxxxxx"
    datasource_uid: "xxxxxxxxxxxx"
    load_balancer: "my-loadbalancer"

datasource_config:
  cloudwatch_config: &cloudwatch_config
      datasource:
        type: cloudwatch
        uid: "${datasource_uid}"
      intervalMs: 50
      logGroups: []
      matchExact: true
      maxDataPoints: 1500
      metricEditorMode: 0
      metricQueryType: 0
      period: ""
      queryMode: Metrics
      region: default
      sqlExpression: ""

common_rule_config: &common_rule_config
    no_data_state: "KeepLast"
    exec_err_state: "KeepLast"
    interval_seconds: 60
    labels: {}
    is_paused: true
    for: "1m"

common_data_config: &common_data_config
    relative_time_range:
      from: 60
      to: 0

rules:
  - name: "My Alert"
    condition: "B"
    <<: *common_rule_config
    annotations:
      __alertId__: "1"
      __dashboardUid__: "${dashboard_uid}"
      __panelId__: "3"
      message: "${load_balancer}: HTTP Code Counts alert"
    data_blocks:
      - ref_id: "A"
        <<: *common_data_config
        datasource_uid: "${datasource_uid}"
        model:
          <<: *cloudwatch_config
          alias: "2XX"
          dimensions:
            LoadBalancer: "${load_balancer}"
          label: "2XX"
          metricName: "HTTPCode_Target_2XX_Count"
          namespace: "AWS/ApplicationELB"
          refId: "A"
          statistic: "Sum"
      - ref_id: "D"
        <<: *common_data_config
        datasource_uid: "${datasource_uid}"
        model:
          datasource_config: *cloudwatch_config
          alias: "5XX"
          dimensions:
            LoadBalancer: "${load_balancer}"
          label: "5XX"
          metricName: "HTTPCode_Target_5XX_Count"
          namespace: "AWS/ApplicationELB"
          refId: "D"
          statistic: "Sum"
      - ref_id: "B"
        <<: *common_data_config
        datasource_uid: "__expr__"
        model:
          conditions:
            - evaluator:
                params: [20]
                type: "lt"
              operator:
                type: "or"
              query:
                params: ["A"]
              reducer:
                params: []
                type: "avg"
            - evaluator:
                params: [0]
                type: "gt"
              operator:
                type: "or"
              query:
                params: ["D"]
              reducer:
                params: []
                type: "avg"
          intervalMs: 1000
          maxDataPoints: 43200
          refId: "B"
          type: "classic_conditions"

我的 Terraform 模块：

locals {
  alert_rules = yamldecode(file("${path.module}/rules.yaml"))

}

resource "grafana_rule_group" "rule_group_0000" {
  name             = "My Alerts"
  folder_uid       = "xxxxxxxxxxxxxxx"
  interval_seconds = 60

    dynamic "rule" {
    for_each = local.alert_rules["rules"]
    content {
      name      = rule.value["name"]
      condition = rule.value["condition"]

      dynamic "data" {
        for_each = rule.value["data_blocks"]
        content {
          ref_id = data.value["ref_id"]
          relative_time_range {
            from = 60
            to   = 0
          }
          datasource_uid = ""
          model = jsonencode({
            model = data.value["model"]
          })
        }
      }
    }
  }
}

现在数据块中的模型结构不同，我想知道如何引用它，是否可以直接引用整个列表而不需要设置单独的键值。另外，我在 YAML 配置中使用锚点和别名；这可以用吗还是我需要删除它们。

我尝试在动态“数据”中使用动态“模型”，但它不允许我这样做并引发错误。

Answer 1

我在这里看到几个问题。首先是

yamldecode

函数无法解析YAML锚点和别名，您需要重写rules.yaml或使用

pyyaml

之类的工具来预处理文件。

第二个问题是你有多个

rules

和多个

data_blocks

，所以你需要迭代它们。例如，类似：

resource "grafana_rule_group" "rule_group_0000" {
  name             = "My Alerts"
  folder_uid       = "xxxxxxxxxxxxxxx"
  interval_seconds = 60

  dynamic "rule" {
    for_each = local.alert_rules["rules"]
    content {
      name      = rule.value["name"]
      condition = rule.value["condition"]

      dynamic "data" {
        for_each = rule.value["data_blocks"]
        content {