06-lab4.md

LAB PART.4 Autoscaling

• 延伸 PART.3 加上 Autoscaling 來增減 instances
• 用 stress.php 增加 instances 的 loading

LAB 目標

環境概述

• Region：Oregon (us-west-2)

• Network

  • VPC

    • CIDR：10.128.0.0/16

  • Internet Gateways

  • Subnet

    • public-2a
      • CIDR：10.128.11.0/24
    • public-2b
      • CIDR：10.128.21.0/24
    • private-2a
      • CIDR：10.128.12.0/24
    • private-2b
      • CIDR：10.128.22.0/24

  • Internet Gateways

  • NAT Gateway

    • select public-2a
    • select public-2b

  • Route table

    • Public
      • Destination：0.0.0.0/0 , Target：igw
      • Associations 2 public subnet
    • private-2a
      • Destination：0.0.0.0/0 , Target：nat(public-2a)
      • Associations private-2a subnet
    • private-2b
      • Destination：0.0.0.0/0 , Target：nat(public-2b)
      • Associations private-2b subnet

• EC2

  • Security Group

    • public-ssh
      • Port：22, source：MyIP
    • private-ssh
      • Port：22, source：sg-${public-ssh-id}
    • web-lb
      • Port：80, source：0.0.0.0/0
    • web-wb
      • Port：80, source：sg-${web-lb-id}

  • Instances

    • Bastion
      • public subnet
      • assign public ip
    • WEB
      • private subnet
      • don't assign public ip
      • install apache and php
      • install host.php
      • install stress.php
      • build image to ami

  • Load balancer

    • Class LB
      • Security Groups：web-lb
      • Health Check
        • Ping Protocol：HTTP
        • Ping Port：80
        • Ping Path：/host.php
        • Response Timeout：5
        • Interval：10
        • Unhealthy threshold：2
        • Healthy threshold：2

  • Auto scaling

    • Launch Configurations
      • security groups trust port 80 source anywhere
    • Auto Scaling Groups
      • select private-{2a,2b}
      • add elb
      • health check use elb (check use ec2 type)
      • add instance tag
      • Desired：2
      • Min：1
      • Max：5
      • Scaling Policies
        • scale-out
          • Whenever Maximum of CPU
          • Is >= 70%
          • Add 67% percent of group when 70%
        • scale-in
          • Whenever Maximum of CPU
          • Is <= 30%
          • remove 1 instance when 30%

clean up

• EC2
  • ELB
  • AMI
  • AutoScaling
• VPC
  • NAT Gateway
  • EIP