如何在Terraform中合并RDS与只读副本的DB标识符

要同时为主RDS实例和对应的只读副本创建CloudWatch告警，你需要先构建一个包含所有目标实例（主实例+符合条件的只读副本）的集合，再基于这个集合遍历创建告警。具体步骤如下：

1. 重构本地变量，包含所有需监控的实例 #

修改locals块，先合并所有主实例，再生成符合条件的只读副本实例ID，最后将两者合并为完整的监控实例列表：

locals {
  # 合并PostgreSQL和MariaDB的主RDS实例
  primary_rds_instances = merge(var.rds_postgress, var.rds_mariadb)
  
  # 生成需监控的只读副本实例：仅针对read_replica=true的主实例，按规则生成副本ID
  read_replica_rds_instances = {
    for key, value in local.primary_rds_instances : 
    replace("${key}-read-replica", " ", "") => value 
    if value.read_replica == true
  }
  
  # 合并主实例与只读副本，得到所有需要创建告警的RDS实例
  all_rds_instances = merge(local.primary_rds_instances, local.read_replica_rds_instances)
}

2. 修改CloudWatch告警资源的遍历目标 #

将告警资源的for_each指向新的local.all_rds_instances，这样就能同时为主实例和只读副本创建告警：

resource "aws_cloudwatch_metric_alarm" "rds_cpu" {
  for_each = local.all_rds_instances

  alarm_name                = replace("${each.key}-cpu-utilization", " ", "")
  comparison_operator       = "GreaterThanThreshold"
  evaluation_periods        = 1
  metric_name               = "CPUUtilization"
  namespace                 = "AWS/RDS"
  period                    = 900
  statistic                 = "Average"
  threshold                 = 80
  alarm_description         = "监控RDS实例CPU使用率"
  insufficient_data_actions = []

  dimensions = {
    DBInstanceIdentifier = each.key
  }
}

说明 #

• read_replica_rds_instances通过for表达式筛选出read_replica=true的主实例，按照你指定的规则生成只读副本的DBInstanceIdentifier作为键
• merge操作将主实例和只读副本实例合并为一个统一的集合，确保所有目标实例都被覆盖
• 告警的dimensions直接使用each.key，对应实例的实际ID，无论是主实例还是只读副本都能正确匹配CloudWatch指标

内容的提问来源于stack exchange，提问作者colbydh