KodeKloud Notes

In this walkthrough, you’ll create your first AWS Fault Injection Simulator (FIS) experiment against the PET Adoption website. By using the AWS FIS API to simulate a disk-fill failure on an EC2 instance, you’ll learn how your application responds to storage exhaustion in one Availability Zone (AZ). This test is crucial for validating your system’s resilience.

Prerequisites

AWS CLI v2 configured with appropriate IAM permissions for FIS and EC2
Target PET Adoption environment deployed across multiple AZs
AWS FIS service enabled in your chosen region

FIS Experiment Components

Before running the scenario, define the two core elements of any chaos experiment:

Given
The current architecture of the PET Adoption website includes:
- EC2 instances distributed across multiple Availability Zones
- An Elastic Load Balancer handling incoming traffic
- An Auto Scaling Group maintaining desired instance count
Resource Role
EC2 instances (multi-AZ) Application web servers
Elastic Load Balancer (ELB) Distributes traffic evenly
Auto Scaling Group (ASG) Scales instances based on demand
Hypothesis
Injecting a disk-fill failure on a single EC2 instance in one AZ will not impact overall availability. With traffic routed to healthy instances in other AZs, the PET Adoption web page should remain responsive.

Resource	Role
EC2 instances (multi-AZ)	Application web servers
Elastic Load Balancer (ELB)	Distributes traffic evenly
Auto Scaling Group (ASG)	Scales instances based on demand

The image illustrates a disk fill scenario on AWS EC2, showing a Virtual Private Cloud (VPC) setup with various components and a hypothesis that an EC2 disk filling up in a single availability zone will not affect application availability.

Next, we’ll walk through the step-by-step execution of this FIS experiment and verify whether our hypothesis holds true.

References

Watch Video

Watch video content