Creating fault-tolerant System with ELB, Auto-scale & CloudWatch Alarm

In this lab we will create fault system in which whenever of the server behind load balancer goes down, cloudwatch alarm should trigger auto-scaling to add one more behind it

• Create AMI for Auto-scaling launch configuration by following Lab: 4.11 & 4.13
• Create a CLB by following Lab 5.2
• Now let’s start creating auto-scaling, please follow along with screenshots
  
• Now select the AMI created in starting steps
  

• Now create Auto-scaling Group
  

• Select your Auto-scaling group and edit
  
• Change max to 4, so more instances can be added to this Auto-scaling group whenever scaling policies are triggered
  Also, add CLB created in starting steps so whatever EC2 is created from this ASG is added behind CLB
  
• Now go to cloudwatch > Alarms > Create Alarm
  
• Select metric
  
• Select ELB
  
• Select By Namespace
  
• Select HealthHostCount, as we will monitoring no. of health hosts behind CLB
  
• So this alarm should trigger whenever no. of host behind CLB is less then 2
  
• Now go to Auto Scaling Groups > Scaling Policies > Add Policy
  
• Select Create a simple policy
  
• Name this policy and select the alarm created in above steps
  ASG should one instance whenever this policy is triggered
  Create!
  
• To test this architecture, just login into anyone of the instances created by auto-scaling and remove the health.html on which CLB is performing health check this should simulate the failure and cloudwatch alarm should get trigger and in turn cloudwatch alarm should trigger ASG scaling policy which should one instance and add it behind CLB
  
• Voila!
  
• Delete everything ;(