I've seen an example of how to trigger an autoscale-group to cloudwatch, but what is the correct method of sending SNS notification of ELB health degradation? I have created a resource:
resource "aws_cloudwatch_metric_alarm" "elb_warn" {
alarm_name = "awselb-${var.service}-Warning-Healthy-Hosts"
comparison_operator = "LessThanThreshold"
evaluation_periods = "4"
metric_name = "HealthyHostCount"
namespace = "AWS/ELB"
period = "3600"
statistic = "Average"
threshold = "2.5"
alarm_description = "This metric alerts elb health warning for ${var.service}"
alarm_actions = [ "what to put here?" ]
}