使用Terraform自动设置云监控警报用于自动扩展的服务器

aws-scripts-mon

我们将使用Terraform为注册在CloudWatch中的数据设置警报。

使用aws-scripts-mon时，通过将–autoscaling与dimensions进行组合，可以将AutoScalingGroupName添加到其中。

mon-put-instance-data.pl --mem-util --disk-path=/ --disk-space-util --auto-scaling --from-cron

我将在通过autoscaling启动的服务器上注册一个cron。

需要注意的是，如果IAM策略中没有ec2:DescribeTags权限，将无法获取AutoscalingGroupName并添加到dimensions中。

自动伸缩

只要按照以下步骤，就可以进行CloudWatch配置。假设已经创建了aws_sns_topic.notify_to_slack.arn和aws_autoscaling_group.front。

请注意，我的原生语言是英语，所以可能无法提供完全准确的翻译。

resource "aws_cloudwatch_metric_alarm" "front_cpu" {
    alarm_name = "${aws_autoscaling_group.front.name} CPU Utilization"
    comparison_operator = "GreaterThanOrEqualToThreshold"
    evaluation_periods = "2"
    metric_name = "CPUUtilization"
    namespace = "AWS/EC2"
    dimensions { "AutoScalingGroupName"="${aws_autoscaling_group.front.name}" }
    statistic = "Average"
    period = "300"
    threshold = "80"
    alarm_description = "${aws_autoscaling_group.front.name} CPU Utilization"
    alarm_actions = [ "${aws_sns_topic.notify_to_slack.arn}" ]
    insufficient_data_actions = [ "${aws_sns_topic.notify_to_slack.arn}" ]
    ok_actions = [ "${aws_sns_topic.notify_to_slack.arn}" ]
}

resource "aws_cloudwatch_metric_alarm" "front_mem" {
    alarm_name = "${aws_autoscaling_group.front.name} Memory Utilization"
    comparison_operator = "GreaterThanOrEqualToThreshold"
    evaluation_periods = "2"
    metric_name = "MemoryUtilization"
    namespace = "System/Linux"
    dimensions { "AutoScalingGroupName"="${aws_autoscaling_group.front.name}" }
    statistic = "Average"
    period = "300"
    threshold = "90"
    alarm_description = "${aws_autoscaling_group.front.name} Memory Utilization"
    alarm_actions = [ "${aws_sns_topic.notify_to_slack.arn}" ]
    insufficient_data_actions = [ "${aws_sns_topic.notify_to_slack.arn}" ]
    ok_actions = [ "${aws_sns_topic.notify_to_slack.arn}" ]
}

例子

通常情况下，对于实例我们使用InstanceId作为dimensions。

如果在创建 aws_instance 时使用了 count，那么您也可以在 aws_cloudwatch_metric_alarm 中使用 count 来进行相应的处理。

resource "aws_cloudwatch_metric_alarm" "elasticsearch_cpu" {
    alarm_name = "${element(aws_instance.elasticsearch.*.id, count.index)} CPU Utilization"
    comparison_operator = "GreaterThanOrEqualToThreshold"
    evaluation_periods = "2"
    metric_name = "CPUUtilization"
    namespace = "AWS/EC2"
    dimensions { "InstanceId"="${element(aws_instance.elasticsearch.*.id, count.index)}" }
    statistic = "Average"
    period = "300"
    threshold = "80"
    alarm_description = "${element(aws_instance.elasticsearch.*.id, count.index)} CPU Utilization"
    alarm_actions = [ "${aws_sns_topic.notify_to_slack.arn}" ]
    insufficient_data_actions = [ "${aws_sns_topic.notify_to_slack.arn}" ]
    ok_actions = [ "${aws_sns_topic.notify_to_slack.arn}" ]
    count = "${var.elasticsearch_server_count}"
}
resource "aws_cloudwatch_metric_alarm" "elasticsearch_mem" {
    alarm_name = "${element(aws_instance.elasticsearch.*.id, count.index)} Memory Usage"
    comparison_operator = "GreaterThanOrEqualToThreshold"
    evaluation_periods = "2"
    metric_name = "MemoryUtilization"
    namespace = "System/Linux"
    dimensions { "InstanceId"="${element(aws_instance.elasticsearch.*.id, count.index)}" }
    statistic = "Average"
    period = "300"
    threshold = "90"
    alarm_description = "${element(aws_instance.elasticsearch.*.id, count.index)} CPU Utilization"
    alarm_actions = [ "${aws_sns_topic.notify_to_slack.arn}" ]
    insufficient_data_actions = [ "${aws_sns_topic.notify_to_slack.arn}" ]
    ok_actions = [ "${aws_sns_topic.notify_to_slack.arn}" ]
    count = "${var.elasticsearch_server_count}"
}