AWS auto-scaling group helps you ensure that you have the correct number of Amazon EC2 instances available to handle the load for your application. By use of  auto-scaling policy, Auto Scaling group can launch or terminate instances as demand on your application increases or decreases.

Today, I will show you how to use Terraform template to setup an AWS auto-scaling group with ELB. My Terraform version is terraform_0.8.8.

My Terraform template includes:

1. Create a aws_launch_configuration (webcluster) which defines how each EC2 instance will be built for an auto-scaling group;
2. Create an AWS auto-scaling group (scalegroup);
3. Create 1st AWS autoscaling policy (autopolicy) for auto-scaling group scale out;
4. Create 2nd AWS autoscaling policy (autopolicy-down) for auto-scaling group scale in;
5. Create 1st AWS CloudWatch alarm (cpualarm) to trigger auto-scaling group to scale out;
6. Create 2nd AWS CloudWatch alarm (cpualarm-down) to trigger auto-scaling group to scale in;
7. Create a security group (websg) to allow HTTP and management SSH connectivity;
8. Create an Elastic load balancer with cookie session persistence and use this load balancer in front of auto-scaling group (scalegroup). The ELB will health check all EC2 instances in the auto-scaling group. If any EC2 instance fails the ELB health check, it won’t receive any incoming traffic. If the existing EC2 instances are overloaed (in our case CPU utilisation is over 60%),  the auto-scaling group will create more EC2 instance to handle the spike. On the contrary, the auto-scaling group will scale in when EC2 instance is idle (CPU utilisation is less than 10%).
9. Create a SSH key pair and use for AWS auto-scaling group (scalegroup);
10. Create output of ELB DNS;

Template

provider “aws” {
region = “ap-southeast-2”
shared_credentials_file = “${pathexpand(“~/.aws/credentials”)}”
#shared_credentials_file = “/home/dzhang/.aws/credentials”
}

resource “aws_launch_configuration” “webcluster” {
image_id= “ami-4ba3a328”
instance_type = “t2.micro”
security_groups = [“${aws_security_group.websg.id}”]
key_name = “${aws_key_pair.myawskeypair.key_name}”
user_data = <<-EOF
#!/bin/bash
echo “hello, I am WebServer” >index.html
nohup busybox httpd -f -p 80 &
EOF

lifecycle {
create_before_destroy = true
}
}

resource “aws_key_pair” “myawskeypair” {
key_name = “myawskeypair”
public_key = “${file(“awskey.pub”)}”
}

data “aws_availability_zones” “allzones” {}

resource “aws_autoscaling_group” “scalegroup” {
launch_configuration = “${aws_launch_configuration.webcluster.name}”
availability_zones = [“${data.aws_availability_zones.allzones.names}”]
min_size = 1
max_size = 4
enabled_metrics = [“GroupMinSize”, “GroupMaxSize”, “GroupDesiredCapacity”, “GroupInServiceInstances”, “GroupTotalInstances”]
metrics_granularity=”1Minute”
load_balancers= [“${aws_elb.elb1.id}”]
health_check_type=”ELB”
tag {
key = “Name”
value = “terraform-asg-example”
propagate_at_launch = true
}
}
resource “aws_autoscaling_policy” “autopolicy” {
name = “terraform-autoplicy”
scaling_adjustment = 1
adjustment_type = “ChangeInCapacity”
cooldown = 300
autoscaling_group_name = “${aws_autoscaling_group.scalegroup.name}”
}

resource “aws_cloudwatch_metric_alarm” “cpualarm” {
alarm_name = “terraform-alarm”
comparison_operator = “GreaterThanOrEqualToThreshold”
evaluation_periods = “2”
metric_name = “CPUUtilization”
namespace = “AWS/EC2”
period = “120”
statistic = “Average”
threshold = “60”

dimensions {
AutoScalingGroupName = “${aws_autoscaling_group.scalegroup.name}”
}

alarm_description = “This metric monitor EC2 instance cpu utilization”
alarm_actions = [“${aws_autoscaling_policy.autopolicy.arn}”]
}

#
resource “aws_autoscaling_policy” “autopolicy-down” {
name = “terraform-autoplicy-down”
scaling_adjustment = -1
adjustment_type = “ChangeInCapacity”
cooldown = 300
autoscaling_group_name = “${aws_autoscaling_group.scalegroup.name}”
}

resource “aws_cloudwatch_metric_alarm” “cpualarm-down” {
alarm_name = “terraform-alarm-down”
comparison_operator = “LessThanOrEqualToThreshold”
evaluation_periods = “2”
metric_name = “CPUUtilization”
namespace = “AWS/EC2”
period = “120”
statistic = “Average”
threshold = “10”

dimensions {
AutoScalingGroupName = “${aws_autoscaling_group.scalegroup.name}”
}

alarm_description = “This metric monitor EC2 instance cpu utilization”
alarm_actions = [“${aws_autoscaling_policy.autopolicy-down.arn}”]
}

resource “aws_security_group” “websg” {
name = “security_group_for_web_server”
ingress {
from_port = 80
to_port = 80
protocol = “tcp”
cidr_blocks = [“0.0.0.0/0”]
}

lifecycle {
create_before_destroy = true
}
}

resource “aws_security_group_rule” “ssh” {
security_group_id = “${aws_security_group.websg.id}”
type = “ingress”
from_port = 22
to_port = 22
protocol = “tcp”
cidr_blocks = [“60.242.xxx.xxx/32”]
}

resource “aws_security_group” “elbsg” {
name = “security_group_for_elb”
ingress {
from_port = 80
to_port = 80
protocol = “tcp”
cidr_blocks = [“0.0.0.0/0”]
}

egress {
from_port = 0
to_port = 0
protocol = “-1”
cidr_blocks = [“0.0.0.0/0”]
}

lifecycle {
create_before_destroy = true
}
}

resource “aws_elb” “elb1” {
name = “terraform-elb”
availability_zones = [“${data.aws_availability_zones.allzones.names}”]
security_groups = [“${aws_security_group.elbsg.id}”]
access_logs {
bucket = “elb-log.davidwzhang.com”
bucket_prefix = “elb”
interval = 5
}
listener {
instance_port = 80
instance_protocol = “http”
lb_port = 80
lb_protocol = “http”
}
health_check {
healthy_threshold = 2
unhealthy_threshold = 2
timeout = 3
target = “HTTP:80/”
interval = 30
}

cross_zone_load_balancing = true
idle_timeout = 400
connection_draining = true
connection_draining_timeout = 400

tags {
Name = “terraform-elb”
}
}

resource “aws_lb_cookie_stickiness_policy” “cookie_stickness” {
name = “cookiestickness”
load_balancer = “${aws_elb.elb1.id}”
lb_port = 80
cookie_expiration_period = 600
}

output “availabilityzones” {
value = [“${data.aws_availability_zones.allzones.names}”]
}

output “elb-dns” {
value = “${aws_elb.elb1.dns_name}”
}

Output:

LauchConfiguration

CloudWatchAlarm

Auto-scaling Policy

Scale Out

Scale In

Auto-scaling group

ELB

EC2 Instance