Ease using AWS ALBs in CloudWatch metrics

[sj26]

Fantastic to see ALBs available in Terraform so quickly. Thank you! 👏

Part of cutting over to ALBs for us involves creating equivalent CloudWatch alarms around our upstream health. The dimensions aren't just the LoadBalancerName set to aws_elb.*.name like ELBs however. For ALBs the dimension is LoadBalancer with a partial suffix from the ALB ARN:

# Alarm if this alb's targets have more than 50 5xx errors within two minutes
resource "aws_cloudwatch_metric_alarm" "my-alb-too-many-server-errors" {
  alarm_name  = "my-alb-too-many-server-errors"
  namespace   = "AWS/ApplicationELB"
  metric_name = "HTTPCode_Target_5XX_Count"

  dimensions = {
    # CloudWatch wants part of the arn. It's kinda silly.
    # arn:aws:elasticloadbalancing:us-east-1:123456:loadbalancer/app/my-alb/abc123 => app/my-alb/abc123
    LoadBalancer = "${replace("${aws_alb.my-alb.arn}", "/arn:.*?:loadbalancer\\/(.*)/", "$1")}"
  }

  statistic           = "Sum"
  period              = 60
  comparison_operator = "GreaterThanThreshold"
  threshold           = "50"
  evaluation_periods  = 2
}

This definitely works, but it's pretty clumsy. Perhaps this ARN suffix could be exposed as a derived attribute of aws_alb resources?

[jen20]

Hi @sj26! This looks like a great candidate for an additional computed field. I'll look into adding this.

[sj26]

Thanks so much! 💖

[sj26]

Any chance we could do the same thing for target group resources?

http://docs.aws.amazon.com/AmazonCloudWatch/latest/monitoring/elb-metricscollected.html#load-balancer-metric-dimensions-alb

[adhodgson1]

It would be good to have this exposed on the target group as well, as it is needed for metrics on that such as the unhealthy host count.

[jaydp17]

For anyone coming here looking for a solution, the below works

• aws_alb.my-alb.arn_suffix
• aws_alb_target_group.my-tg.arn_suffix

[hdryx]

I was searching for this for hours. Why it's not documented here https://www.terraform.io/docs/providers/aws/d/lb.html ?

[ghost]

I'm going to lock this issue because it has been closed for 30 days ⏳. This helps our maintainers find and focus on the active issues.

If you have found a problem that seems similar to this, please open a new issue and complete the issue template so we can capture all the details necessary to investigate further.