C3 - Part II - Monitor and scale your cloud

This page describes the Part II of the Challenge 2

Use Case

Case 1 : As soon as the CPU load of the main Drupal instance reach 70%, the autoscaling mechanism must launch new instances (horizontal scaling) to absorb this load (max 4 in total).

Case 2 : As soon as the load goes down, we must observe the instances reduce in number until the return to normal operation status.

Step 0 - Read the doc

• AWS - What is EC2 autoscaling ?

Step 1 - Create a launch template

• AWS - Tutorial

• Naming convention: LCONFIG_CLGDRP[XX]

• Tips:

  • Use the same SG, same KeyPair as for your original instance.

  • To not use "Spot" Instance.

Step 2 - Create a single-instance Auto Scaling group

• AWS - Tutorial

• Naming convention : ASG_CLDGRP_[XX]_LABO3

• Tips:

  • Step 3

    • Load balancing : Select "No load balancer"

Step 3 - Experiment with your config

• Using Stress as in Part I, try to activate your autoscaling group and observe the instances automatically launched.

• After Stress sequence, observe how your infra return to "normal" operation.