- Application Load Balancer
In this lab I will create an Auto Scaling Group and place it behind an Application Load Balancer. Instances will be deployed within a Target Group. Each Instance will run an Apache server and will be stressed to cause spike in CPU utilization above threshold. That will cause Instances to fail Health Checks which will trigger replacement of the unhealthy one with a new one.
- Create and configure an Application Load Balancer
- Create and configure Auto Scaling groups and launch templates
- Utilize Auto Scaling to ensure application availability
- Stress test Instances using stress tool
- AWS account
- Start with creating a Load Balancer of type ALB. It’ll be Internet-facing and IPv4. Create a new Security Group that will allow all inbound HTTP traffic from anywhere.Back in ALB creation tab attach that SG to Load Balancer.
- In the Listeners and routing click the Create target group link. Use the default values and give it a name. Back in ALB tab choose the TG. Listener should be set as HTTP on port 80. Then Create load balancer.
- Create a Security Group for instances. This group will need inbound rules for both SSH and HTTP from anywhere since it’ll be attached to instances for a webserver cluster.
- Create Launch Template. As AMI choose Amazon LInux 2, Instance type t2.micro, choose or create a Key pair. Then attach the earlier created security group. Then in Advanced details enable Detailed CloudWatch monitoring.
The provided user data will install PHP, an Apache webserver and a tool for stress testing.
#!/bin/bash sudo amazon-linux-extras install epel sudo yum install -y httpd php service httpd start sudo yum install -y stress
- Create an Auto Scaling Group. As a Launch template choose the one created in previous step. For the Network options choose Subnet available in drop-down menu. On next page Attach to an existing load balancer choose the one created earlier. Add optional Health checks on ELB every 120 seconds.
Enable group metrics collection within CloudWatch. On the next page change the Maximum capacity value to 5. Scaling policies – optional select Target tracking scaling policy with Target value for Average CPU utilization of value 80.
Continue further and Create Auto Scaling group. This will launch one Instance.
- Continue to Load Balancer and copy the DNS name in browser
- In this step launch stress tool to observe the Auto scaling group scale-out. Connect to the Instance using EC2 Instance Connect. To start a stress test that will increase CPU utilization run:
stress --cpu 2 --io 1 --vm 1 --vm-bytes 128M --timeout 5m
Go back to Instance and choose Monitoring, choose View in metrics which will open CloudWatch. Change Period to 1 minute. The CPU utilization will spike and after 2 minutes the instance will fail a Health Check and a new one will replace it.