基于Cloudwatch规则创建PagerDuty警报的最佳方法

时间:2019-01-30 20:31:37

标签: terraform amazon-cloudwatch pagerduty

PagerDuty(PD)与Cloudwatch(CW)集成在一起,只要触发CW警报,我就使用它来进行分页:https://support.pagerduty.com/docs/aws-cloudwatch-integration-guide

如果触发了CW规则,我想被寻呼。看来我可以使用PD全局事件路由,然后配置CW输入以发送PD全局事件端点期望的响应。但是我喜欢CW警报发布到SNS主题的方式,我所要做的就是将SNS主题消息转发给PD,因此,如果CW规则中有类似内容,那就太好了。

1 个答案:

答案 0 :(得分:0)

结果是,我可以使用TriggeredRules指标从规则创建CW警报。然后,我可以使用现有的PagerDuty CW集成。这是我编写的Terraform代码:

data "template_file" "ecs_task_stopped" {
  template = <<EOF
{
  "source": ["aws.ecs"],
  "detail-type": ["ECS Task State Change"],
  "detail": {
    "clusterArn": ["arn:aws:ecs:$${aws_region}:$${account_id}:cluster/$${cluster}"],
    "desiredStatus": ["Running"],
    "lastStatus": ["STOPPED"]
  }
}
EOF

  vars {
    account_id = "${data.aws_caller_identity.current.account_id}"
    cluster    = "${var.ecs_cluster_name}"
    aws_region = "${data.aws_region.current.name}"
  }
}

resource "aws_cloudwatch_event_rule" "ecs_task_stopped" {
  count         = "${var.should_create == "true" ? 1 : 0}"
  name          = "${var.env}_${var.ecs_cluster_name}_task_stopped"
  description   = "${var.env}_${var.ecs_cluster_name} Essential container in task exited"
  event_pattern = "${data.template_file.ecs_task_stopped.rendered}"
}

resource "aws_cloudwatch_metric_alarm" "alarm_task_stopped_rule_triggered" {
  count               = "${var.should_create == "true" ? 1 : 0}"
  alarm_name          = "${var.ecs_cluster_name}-task-stopped"
  comparison_operator = "GreaterThanOrEqualToThreshold"
  evaluation_periods  = "1"
  datapoints_to_alarm = "1"
  metric_name         = "TriggeredRules"
  namespace           = "AWS/Events"
  period              = "60"
  statistic           = "Maximum"
  threshold           = "1"
  alarm_description   = "Essential container in ${var.ecs_cluster_name} task exited"
  alarm_actions       = ["${var.cw_sns_topic_id}"]
  ok_actions          = ["${var.cw_sns_topic_id}"]

  dimensions {
    RuleName = "${aws_cloudwatch_event_rule.ecs_task_stopped.name}"
  }
}