ECS CloudWatch

Set the variables

The enable_alerts variable is used to turn on/off notifications per environment.

variable "enable_alerts" {  type        = bool  description = "When enabled CloudWatch alarm events are sent to the Alerts SNS Topic"  default     = false}variable "api_service_minimum_task_count" {  type        = number  description = "Minimum number of expected tasks to be running for the API Service"  default     = 1}

Alarms

Each ECS service has an expected number of tasks running, often this is just a single task (e.g the web server). It's a useful way to check if the ECS service is healthy and alert on it.

info

Your Docker image should also have a health check and the service itself a health check endpoint which you can hook Route53 health checks into.

resource "aws_cloudwatch_metric_alarm" "ecs_webapp_task_count_too_low" {  alarm_name          = "ecs-${aws_ecs_service.webapp.name}-lowTaskCount"  comparison_operator = "LessThanThreshold"  evaluation_periods  = 1  metric_name         = "CPUUtilization"  namespace           = "AWS/ECS"  period              = 60  statistic           = "SampleCount"  threshold           = var.webapp_minimum_task_count  treat_missing_data  = "breaching"  alarm_description   = "Task count is too low."  alarm_actions       = var.enable_alerts == true ? [aws_sns_topic.sns_technical_alerts.arn] : []  ok_actions          = var.enable_alerts == true ? [aws_sns_topic.sns_technical_alerts.arn] : []
  dimensions = {    ClusterName = "${aws_ecs_cluster.main.name}"    ServiceName = "${aws_ecs_service.webapp.name}"
  }}
resource "aws_cloudwatch_metric_alarm" "ecs_webapp_cpu_too_high" {  alarm_name          = "ecs-${aws_ecs_service.webapp.name}-highCPUUtilization"  comparison_operator = "GreaterThanThreshold"  evaluation_periods  = 5  metric_name         = "CPUUtilization"  namespace           = "AWS/ECS"  period              = 300  statistic           = "Average"  threshold           = 80  treat_missing_data  = "breaching"  alarm_description   = "Average CPU utilization is too high."  alarm_actions       = var.enable_alerts == true ? [aws_sns_topic.sns_technical_alerts.arn] : []  ok_actions          = var.enable_alerts == true ? [aws_sns_topic.sns_technical_alerts.arn] : []
  dimensions = {    ClusterName = "${aws_ecs_cluster.main.name}"    ServiceName = "${aws_ecs_service.webapp.name}"  }}
resource "aws_cloudwatch_metric_alarm" "ecs_webapp_memory_too_high" {  alarm_name          = "ecs-${aws_ecs_service.webapp.name}-highMemoryUtilization"  comparison_operator = "GreaterThanThreshold"  evaluation_periods  = 5  metric_name         = "MemoryUtilization"  namespace           = "AWS/ECS"  period              = 300  statistic           = "Average"  threshold           = 80  treat_missing_data  = "breaching"  alarm_description   = "Average Memory utilization is too high."  alarm_actions       = var.enable_alerts == true ? [aws_sns_topic.sns_technical_alerts.arn] : []  ok_actions          = var.enable_alerts == true ? [aws_sns_topic.sns_technical_alerts.arn] : []
  dimensions = {    ClusterName = "${aws_ecs_cluster.main.name}"    ServiceName = "${aws_ecs_service.webapp.name}"  }}